Canadian employment trends in the era of generative artificial intelligence: Early evidence

Release date: January 28, 2026

Abstract

Artificial intelligence (AI) holds the potential to transform the nature of work, and its ability to replace human labour remains a central concern. This study highlights recent labour market trends, distinguishing jobs potentially more exposed to and less complementary with AI from other jobs. From November 2022—when generative AI applications started gaining traction following the mass availability of ChatGPT—to December 2025, employment generally grew regardless of potential occupational exposure to and complementarity with AI. However, job growth varied across worker characteristics. Younger employees and those less educated generally saw weaker job growth over this period. Coding-intensive professions (e.g., software engineers and web designers) grew at a similar rate as other jobs. However, gains in coding-intensive jobs were concentrated among workers aged 30 to 49, while the number of coding professionals younger than 30 stagnated. From the fourth quarter of 2022 to the third quarter of 2025, job vacancies in occupations potentially more exposed to and less complementary with AI decreased at a similar rate as vacancies in occupations potentially less exposed to AI. Jobs potentially more exposed to AI regardless of complementarity are more likely to be higher-paying, associated with workplace pension plans, full-time and permanent. Thus, AI-driven layoffs could potentially involve the loss of high-quality jobs. Some of the results reflect longer-term trends predating the widespread availability of AI. It is unclear whether more recent trends reflect the advent of AI, other economic factors such as labour market adjustments after the COVID-19 pandemic, rapid demographic shifts, recent trade tensions with the United States or a combination of factors that are shaping the Canadian economic landscape.

Authors

Tahsin Mehdi and Marc Frenette are with the Economic and Social Analysis and Modelling Division, Analytical Studies and Modelling Branch, at Statistics Canada.

Acknowledgments

The authors would like to thank Ryan Macdonald, Ping Ching Winnie Chan, André Bernard and René Morissette (retired) from Statistics Canada, as well as Gabriela Galassi from the Bank of Canada, for their helpful comments and suggestions.

Introduction

Artificial intelligence (AI)—particularly generative AI—has emerged as a new and potentially transformative force restructuring the labour market.^Note  As these tools become more integrated into different parts of the economy, questions are arising about their potential impact on employment, job quality and the nature of work itself. The Canadian labour market has gone through substantial changes over the past few years because of a range of factors, including the COVID-19 pandemic, population growth driven by immigration and trade tensions with the United States beginning in early 2025. Therefore, isolating the impact of AI from other economic factors is challenging. Nonetheless, tracking and monitoring labour market trends against the backdrop of technological progress are important for policy makers, employers and workers alike as Canada and the rest of the world navigate the potential challenges and opportunities that may lie ahead.

Ever since the seminal work of Frey and Osborne (2013), a growing body of literature on the potential impact of technological transformation on the labour market has emerged (e.g., Oschinski and Wyonch, 2017; Nedelkoska and Quintini, 2018; Frenette and Frank, 2020; Felten et al., 2021; Pizzinelli et al., 2023; Mehdi and Morissette, 2024; Eloundou et al., 2024; Kochhar, 2024; and Gmyrek et al., 2025). Unlike previous waves of technological transformation, such as automation, which primarily affects workers who perform simple and repetitive tasks, AI holds the potential to transform the jobs of workers who perform complex and cognitive tasks.

The percentage of Canadian businesses that reported using AI to produce goods or deliver services doubled from 6% during the 2023-to-2024 period to 12% during the 2024-to-2025 period (Bryan et al., 2024; Bryan et al., 2025; Statistics Canada, 2024a, 2024b, 2025a, 2025b). However, the deployment of AI does not necessarily translate to job loss. Indeed, the percentage of AI-adopting businesses that reported reducing employment because of AI remained steady at about 6% over both periods. But the number and characteristics of workers displaced by AI are unknown. Moreover, these statistics capture direct firm-level employment changes and do not consider the potential indirect effects of AI adoption on other businesses or industries. While this study makes no attempt to quantify AI-driven job losses (direct or indirect), it highlights recent employment trends distinguishing occupations potentially more exposed to and less complementary with AI from other occupations.

AI technologies are rapidly evolving, but their net economic impact remains ambiguous. While AI is capable of potentially replacing a wide range of tasks, such as writing, coding, translating and generating images, it can also augment jobs and have a complementary effect for some workers.^Note  Mehdi and Morissette (2024) and Mehdi and Frenette (2024) estimated that 60% of the Canadian workforce was potentially highly exposed to AI-related job transformation, but AI may be able to augment the jobs of about half of these workers, rather than replace them. Li and Dobbs (2025) applied the same methodology used by Mehdi and Morissette (2024) and Mehdi and Frenette (2024) to online job posting data—they found no clear evidence of AI-driven skills polarization on the labour demand side. Oschinski and Walia (2025) used a large language model-driven approach and found that AI may be more likely to transform the nature of work itself rather than substitute for human labour. To date, no work in Canada has tracked the state of jobs that could potentially be more exposed to AI-related job transformation. Some American studies show that recent employment trends for occupations potentially more exposed to AI-related job transformation are not all that different from those for other occupations (e.g., Eckhardt and Goldschlag, 2025; Gimbel et al., 2025; Hampole et al., 2025; and Humlum and Vestergaard, 2025). Other studies found some correlation between the rise of generative AI and weaker employment growth for younger software developers since the release of ChatGPT in November 2022 (e.g., Brynjolfsson et al., 2025; and Chandar, 2025).

However, it is unclear whether such trends can be attributed to AI alone. Job vacancies surged in late 2021 and continued to increase into 2022 as the economy was recovering from the pandemic, but this unmet demand for skills subsequently subsided (Statistics Canada, 2025c). Thus, it is difficult to ascertain whether any recent employment declines are a result of AI, some post-pandemic labour market adjustment process or a combination of factors beyond these. That said, the adoption and diffusion of new technologies typically unfold gradually, and it may be too early to detect large-scale effects of AI on the labour market. A few years of post-ChatGPT data points should therefore be interpreted with caution, as they may not be indicative of longer-term trends, especially given the concurrent influence of other economic shocks resulting from rapid demographic changes because of increased immigration and recent trade tensions with the United States.

The estimates presented in this study are largely based on the technological feasibility of replacing job tasks. Employers may not immediately replace human labour with AI, even if it is technologically feasible to do so, because of financial, legal and institutional constraints. Consequently, exposure to AI does not necessarily imply a risk of job loss. At the very least, it could imply a certain degree of job transformation.

Data and methods

This study pulls together data from the monthly Labour Force Survey (LFS) to examine employment and the quarterly Job Vacancy and Wage Survey (JVWS) to examine unmet skills demand. The LFS sample is restricted to employees (i.e., paid workers) aged 15 and older.^Note  Both the LFS and the JVWS samples are restricted to the 10 provinces. This study examines labour market trends, leveraging over a decade of data from the LFS (January 2015 to December 2025) and the JVWS (the first quarter of 2015 to the third quarter of 2025).^Note 

There is no straightforward way to measure technology-driven job losses. As a result, analysts often rely on expert assessments to classify occupations by their degree of potential exposure to technological transformation. By comparing employment trends across these exposure gradients, analysts can gain insight into how the deployment of new technologies, such as AI, may be affecting different segments of the labour market, even without explicit data on technology-driven job displacement.

Following U.S. studies such as those of Brynjolfsson et al. (2025) and Gimbel et al. (2025), this study examines labour market trends relative to November 2022, when generative AI tools started gaining traction following the mass availability of ChatGPT. While November 2022 marks an important milestone for the widespread availability of generative AI, it is important to recognize that occupational shifts may have already been underway, because earlier advances in machine learning and related technologies likely began reshaping work processes and job requirements well before this date. In Canada specifically, technological developments in generative AI beginning in 2022 also coincided with rapid demographic changes driven largely by increased immigration—especially the substantial rise in the number of international students entering the labour market. Thus, observed changes in the labour market since November 2022 may reflect the cumulative effects of recent generative AI developments, the broader ongoing influence of previous AI technologies and rapid demographic changes.

Two different methods that have been used to group occupations based on potential exposure to AI include (1) the complementarity-adjusted AI occupational exposure (C-AIOE) index developed by Felten et al. (2021) and the International Monetary Fund (Pizzinelli et al., 2023), and (2) the isolation of coding-intensive jobs.^Note 

The C-AIOE index is the main method used in this study. This measure was also used by Mehdi and Morissette (2024), Mehdi and Frenette (2024), and Li and Dobbs (2025) among others. The C-AIOE index measures the degree to which a given occupation may be exposed to AI and the extent to which it could be complementary with AI.

Quantifying potential exposure relies on expert assessments of AI’s capacity to mimic a variety of human occupational abilities, such as writing, reasoning, memorizing and physical strength, and the integration of these assessments with data on the prevalence and importance of these abilities across thousands of jobs measured by the Occupational Information Network (O*NET).^Note 

Quantifying potential complementarity requires work context and skills data from O*NET. Potential complementarity with AI depends on the extent to which a job’s tasks and work context can benefit from AI as a supportive tool, rather than being directly substituted by it. For example, occupations that require advanced communication, public speaking or nuanced face-to-face interactions often rely on human strengths that are not easily replicated by AI, making these roles more likely to be augmented—rather than replaced—by new technologies. Similar patterns hold in jobs where responsibility for outcomes, especially those relating to others’ health or safety, remains crucial; here, human oversight, judgment and accountability still play an essential role even as AI systems assist with analysis or routine processes. Additionally, occupations that necessitate working in complex physical environments or close proximity to others and those where the consequences of errors are significant tend to require human adaptability and real-time decision making, limiting the extent to which AI can fully substitute for human workers. Finally, jobs that are less routine or demand considerable preparation and specialized skills are more likely to harness AI as a tool for augmentation, since integrating and leveraging AI often require a high degree of expertise and adaptability. Thus, the potential for AI to complement, rather than replace, workers varies according to the work context, task structure and skill requirements across occupations. See Felten et al. (2021) and Pizzinelli et al. (2023) for full details on the computation of the C-AIOE measure.

The C-AIOE index can group jobs into three distinct categories based on whether the potential occupational exposure and complementarity scores fall above or below the median score across all occupations: (1) high exposure and low complementarity (HELC), (2) high exposure and high complementarity (HEHC), and (3) low exposure (LE). Thus, these potential occupational exposure and complementarity rankings are relative and not absolute. That is, a job is only more or less potentially exposed to or complementary with AI relative to the median occupational score.

HELC jobs—which comprise a mix of skill levels ranging from retail salespeople, data entry clerks and other office support workers to software engineers, economists, accountants and financial auditors—involve tasks that may be more susceptible to replacement by AI. There is considerable uncertainly, however, regarding the extent to which AI can actually replace human labour. For example, although thought to be highly exposed to and less complementary with AI, software engineers and adjacent jobs are integral to maintaining and improving future AI infrastructure. HEHC jobs—which comprise professions such as doctors, nurses, teachers and engineers—are characterized by advanced training requirements and tasks that AI may be more likely to augment rather than substitute for human labour.^Note  LE jobs—which comprise a wide range of occupations from the skilled trades, cashiers, bartenders and chefs to first responders—are potentially the least exposed to AI, but they may face a relatively higher risk of automation-related job transformation.^Note  Thus, it is important to recognize that recent labour market trends may not always show stark differences in outcomes between HELC and LE jobs, given that both groups may be vulnerable to technological transformation, whether through AI or automation. Moreover, shifts in employment may reflect not only the displacement of jobs, but also the expansion or creation of occupations. Ultimately, the impact of emerging technologies on different occupational groups is complex and multifaceted, with the potential for job loss and job creation depending on the specific context. This suggests that the effects of technological transformation on job displacement may not be uniform or solely confined to the most exposed occupations.

The methods used in this study rely on expert assessments of AI’s current capacity to perform occupational tasks. The applicability of these measures may therefore decrease over time as AI technologies improve. Moreover, the measures capture only the direct potential impact of AI. They do not capture spillover effects such as the creation of new jobs, as was the case with previous technological innovations (e.g., the invention of the personal computer led to job losses but also job creation). It is important to note that, as with any measure of technological transformation—whether related to AI or earlier waves of automation—expert assessments retain an element of subjectivity. Nevertheless, they provide valuable and systematic insight into how different occupations may interact with emerging technologies.

Employment generally grew regardless of potential occupational exposure to and complementarity with artificial intelligence from November 2022 to December 2025

Regardless of potential occupational exposure to and the degree of complementarity with AI, employment generally grew at similar rates from November 2022 to December 2025 (Chart 1). There is no clear evidence of any persistent decline in HELC jobs for men or women over this period. Employment for men in HELC jobs was about 10% higher in December 2025 than in November 2022. For women, it was around 5% higher. These growth rates were not significantly different from those observed for HEHC and LE jobs, and neither were the gender differences across the respective C-AIOE groups. In the longer term, however, employment growth in HELC and HEHC jobs outpaced growth in LE jobs over the last decade for both men and women.^Note 

Employment growth was relatively weak for younger employees and less educated employees from November 2022 to December 2025 regardless of potential occupational exposure to and complementarity with artificial intelligence

Although the immediate impact of generative AI on aggregate employment may not yet be detectable on a large scale, there are growing concerns that it may be displacing certain groups of workers, such as those in entry-level positions. Chart 2 illustrates this issue by comparing employment growth among employees aged 15 to 29 and employees aged 30 to 49 across the three C-AIOE groups.^Note  Comparing employment growth across age groups can be challenging given the rapid demographic shifts observed in Canada in recent years caused by increased immigration beginning in 2022—especially the substantial rise in the number of international students entering the labour market. However, from 2022 to 2025, the annual population growth rates of 15- to 29-year-olds (9%) and 30- to 49-year-olds (11%) were similar, helping to ensure that the observed comparisons in employment growth between these age groups were not confounded by uneven population changes (Statistics Canada, 2025d). Panel (a) shows that employment among employees aged 15 to 29 in December 2025 was roughly 5% higher than in November 2022. Moreover, the differences between HELC job growth and HEHC and LE job growth from November 2022 to December 2025 were not statistically significant (albeit less than a fifth of employees aged 15 to 29 were in HEHC jobs). Employment growth for younger employees across the C-AIOE groups was generally stagnant compared with their older counterparts, who saw more robust growth. Employment for those aged 30 to 49 steadily grew across all three C-AIOE groups during this period. By December 2025, employment among employees aged 30 to 49 was 10% higher than in November 2022. As was the case for younger employees, the differences in growth rates across the three C-AIOE groups were not statistically significant among older employees.

Although highly skilled workers were largely unaffected by previous waves of technological transformation, such as automation, there are growing concerns that even highly skilled jobs may be affected or transformed in some way in the era of generative AI. Employment growth trends across the three C-AIOE groups varied considerably by educational attainment (Chart 3). Employees with a bachelor’s degree or above saw the most robust growth from November 2022 to December 2025. Regardless of potential occupational exposure to and complementarity with AI, they saw job gains of 10% to 20% over this period.^Note  Employees whose highest level of education was a trades certificate or diploma saw job gains of 10% in LE jobs and 20% in HEHC jobs but no growth in HELC jobs. The lack of growth in HELC jobs is not surprising given that the skilled trades are largely concentrated in the LE group. By contrast, there was little to no employment growth regardless of potential occupational exposure to and complementarity with AI for employees with a high school diploma or below or with a postsecondary credential below a bachelor’s degree or some postsecondary education (incomplete). The differences in growth rates across HEHC, HELC and LE jobs within the educational attainment categories were not statistically significant. Some of the divergent employment growth trends across education levels can be explained by occupational composition differences. For example, less educated employees were more likely to be employed in retail services, data entry and other clerical occupations. Data entry and clerical occupations have generally decreased over time. Highly educated employees were more likely to be employed in digital technology-related occupations, which represented a growing share of occupations.

Data table for Chart 1

Data table for chart 1
Table summary
This table displays the results of Data table for chart 1 Men+, Women+, High exposure and high complementarity, High exposure and low complementarity, Low exposure, High exposure and high complementarity, High exposure and low complementarity and Low exposure, calculated using index units of measure (appearing as column headers).

	Men+			Women+
	High exposure and high complementarity	High exposure and low complementarity	Low exposure	High exposure and high complementarity	High exposure and low complementarity	Low exposure
	index
Notes: The dashed vertical line demarcates the beginning of mass availability of ChatGPT in November 2022. The solid horizontal line demarcates the baseline of the employment growth index. Any point above the line indicates a higher employment level relative to the one observed in November 2022 and any point below the line indicates a lower employment level. Employment refers to the main job of employees aged 15 and older. Given that the non-binary population is small, data aggregation to a two-category gender variable is necessary to protect the confidentiality of responses. Individuals in the category “non-binary people” are distributed into the other two gender categories and are denoted by the “+” symbol in published tables and microdata. The category “men+” includes men, boys and some non-binary people, while the category “women+” includes women, girls and some non-binary people. For ease of communication, the text simply refers to men and women. Employment growth is expressed relative to the level observed in November 2022. For example, an index of 110 would indicate 10% growth (110 minus 100) in employment relative to the November 2022 level, and an index of 90 would indicate that employment was 10% lower (90 minus 100) relative to the November 2022 level. Employment was disaggregated based on the complementarity-adjusted artificial intelligence occupational exposure (C-AIOE) index developed by Felten et al. (2021) and Pizzinelli et al. (2023). An occupation has potentially high exposure if its AIOE index exceeds the median AIOE across all occupations; otherwise, it is considered to have low exposure. Similarly, an occupation is considered to have potentially high complementarity if its complementarity parameter exceeds the median complementarity across all occupations; otherwise, it is considered to have low complementarity. Sources: Statistics Canada, Labour Force Survey, January 2015 to December 2025; and Occupational Information Network.
2015
January	76.7	78.9	96.5	79.3	88.4	98.5
February	77.4	78.7	96.3	79.7	88.3	98.1
March	78.9	77.3	96.1	79.8	87.7	99.2
April	79.1	77.0	96.3	79.7	87.6	99.5
May	79.2	78.0	95.9	80.2	86.9	98.9
June	79.8	78.7	96.3	81.9	86.7	98.4
July	80.1	79.3	95.8	81.8	87.3	98.1
August	82.4	78.5	95.6	82.0	87.3	97.9
September	82.5	79.1	95.2	82.1	87.6	97.1
October	82.8	79.5	94.9	80.6	89.1	96.6
November	82.3	78.2	94.8	82.3	88.0	96.7
December	81.8	79.6	94.2	81.6	87.9	96.9
2016
January	81.7	79.3	94.2	81.1	87.9	99.0
February	81.2	80.0	94.3	80.7	88.2	98.4
March	81.0	81.3	93.7	80.7	88.7	98.4
April	81.5	81.7	93.3	81.4	88.7	98.5
May	82.1	81.5	93.7	81.1	89.2	98.5
June	82.3	80.2	93.7	81.6	88.9	98.7
July	82.2	80.9	93.4	81.1	88.3	100.1
August	81.6	81.0	94.5	81.7	89.9	97.9
September	81.2	80.9	95.2	81.5	90.1	98.4
October	80.6	80.4	95.5	81.8	90.3	98.6
November	80.0	81.8	96.0	81.2	90.8	98.8
December	81.0	81.6	96.4	81.3	90.9	98.9
2017
January	81.9	80.6	96.8	81.6	93.0	98.7
February	82.3	80.7	97.1	82.1	90.8	99.4
March	83.6	80.4	97.4	82.3	91.1	99.0
April	82.2	80.7	98.1	82.3	90.6	99.5
May	81.4	81.4	98.3	83.1	90.7	100.4
June	82.0	82.5	98.1	82.3	91.5	100.8
July	82.7	81.8	98.0	83.1	92.2	100.7
August	82.9	83.9	97.6	83.1	90.4	100.6
September	82.9	84.0	97.3	83.3	90.3	101.2
October	83.2	85.1	96.4	82.9	90.3	102.0
November	85.3	82.3	98.2	83.5	90.1	101.6
December	85.7	83.4	98.3	83.9	90.3	102.0
2018
January	85.4	85.1	97.2	84.0	90.4	101.5
February	85.6	84.2	97.4	83.3	90.6	101.1
March	85.2	85.2	97.9	84.6	90.7	101.4
April	84.8	84.7	98.1	84.7	91.3	100.9
May	84.9	84.1	98.3	85.0	91.6	101.4
June	84.0	85.3	99.0	84.3	92.6	101.7
July	82.3	87.3	99.4	83.7	93.1	103.7
August	83.1	84.5	99.8	82.1	93.6	103.2
September	83.4	84.3	100.4	82.5	93.1	102.8
October	83.6	82.7	100.5	83.3	92.5	103.9
November	83.4	85.1	100.5	83.7	92.7	104.5
December	83.5	85.2	101.0	85.1	91.4	103.4
2019
January	84.5	85.6	100.9	85.7	92.1	104.5
February	84.2	87.2	99.9	86.0	92.8	104.7
March	85.1	87.3	100.3	85.7	92.7	103.4
April	86.4	88.6	100.2	86.1	93.6	106.3
May	87.6	88.9	100.3	86.0	94.2	106.1
June	88.0	89.0	100.1	86.9	94.1	105.5
July	87.5	89.5	100.2	87.5	93.1	104.8
August	87.2	89.9	100.2	88.1	93.4	105.1
September	87.2	89.7	99.8	88.4	92.9	104.2
October	86.2	89.9	99.7	88.9	93.9	102.8
November	86.3	90.0	99.3	88.8	94.0	102.2
December	86.0	91.3	99.5	87.2	95.1	103.6
2020
January	85.6	91.3	99.9	87.2	95.6	103.4
February	86.6	91.2	100.5	86.7	95.1	103.6
March	84.8	89.5	92.6	82.5	90.1	87.5
April	81.3	80.8	75.8	77.6	80.6	72.6
May	83.4	83.1	79.1	79.3	81.0	74.5
June	83.7	89.8	88.3	83.3	86.9	84.1
July	86.2	88.6	89.5	85.9	90.0	86.2
August	88.3	90.8	90.9	85.9	92.2	88.1
September	87.7	94.7	93.6	87.0	94.4	89.9
October	88.1	94.6	92.9	87.8	94.5	92.3
November	87.2	95.5	93.7	88.0	96.3	91.2
December	87.8	95.8	93.2	89.1	94.5	90.8
2021
January	87.6	93.6	93.6	87.3	92.8	88.1
February	87.7	96.3	94.3	89.3	94.6	91.2
March	87.6	98.1	95.5	91.3	95.8	92.1
April	90.2	97.9	92.7	90.5	95.2	92.7
May	88.6	99.5	92.5	89.1	95.7	91.9
June	89.5	99.8	95.7	91.5	96.7	95.4
July	90.7	101.3	93.9	92.1	97.5	95.2
August	91.8	102.4	93.9	94.0	96.4	96.1
September	93.0	103.3	93.9	93.3	99.2	95.0
October	93.6	102.7	94.3	92.7	98.9	97.9
November	95.0	102.2	95.3	93.5	99.4	99.5
December	96.8	101.6	96.0	93.7	99.5	100.0
2022
January	97.0	101.0	94.7	94.2	99.2	94.6
February	97.1	101.3	97.1	95.3	100.0	100.7
March	97.6	99.2	98.4	96.7	99.2	100.5
April	96.5	101.7	98.8	98.5	99.0	101.2
May	98.2	101.7	98.6	99.8	100.9	100.3
June	97.5	102.2	99.1	101.3	100.6	98.6
July	96.8	101.6	99.5	99.4	100.4	99.7
August	98.6	102.6	98.7	99.7	100.2	98.9
September	98.5	100.2	99.4	98.1	99.4	99.8
October	99.9	99.8	99.9	100.3	99.2	100.0
November	100.0	100.0	100.0	100.0	100.0	100.0
December	98.9	101.3	100.7	99.7	100.0	101.3
2023
January	98.7	102.1	102.4	100.2	101.2	102.3
February	100.0	101.9	101.8	100.4	101.2	103.7
March	101.4	102.2	101.5	100.6	102.3	102.7
April	103.0	102.2	100.9	100.7	103.5	102.0
May	103.9	102.2	102.1	102.5	103.1	101.7
June	105.3	104.9	101.6	101.8	103.1	102.2
July	106.3	101.4	102.5	101.9	104.5	101.1
August	105.3	101.3	103.2	102.4	103.9	102.0
September	105.2	101.0	103.3	103.1	103.0	101.4
October	104.1	103.4	103.3	104.4	101.5	102.6
November	103.5	104.3	103.8	105.4	102.3	101.5
December	105.0	102.9	104.0	106.6	101.8	101.5
2024
January	106.7	104.8	103.2	107.5	101.3	102.0
February	107.1	105.5	102.7	108.4	101.7	101.3
March	107.7	105.7	102.3	108.4	101.9	101.6
April	107.6	105.8	104.1	108.5	101.6	101.9
May	109.0	105.2	103.9	107.8	100.5	105.1
June	106.6	106.0	104.7	107.0	100.7	104.9
July	105.6	107.2	104.6	109.4	100.0	104.1
August	105.2	106.3	105.8	109.8	101.1	103.3
September	104.0	107.5	105.6	109.2	102.7	102.2
October	105.1	106.2	105.9	108.1	103.2	101.6
November	105.6	107.2	105.2	107.6	104.0	102.3
December	105.7	108.8	105.3	108.3	105.0	100.7
2025
January	104.9	108.5	106.5	108.0	105.1	101.6
February	106.3	109.0	106.4	108.0	103.9	102.8
March	104.3	108.8	106.5	107.8	104.3	102.8
April	103.0	110.0	106.9	107.6	104.9	102.3
May	104.2	111.3	106.3	108.2	105.7	102.0
June	104.4	111.6	107.1	108.7	106.0	101.9
July	105.9	111.0	105.5	108.7	105.3	102.5
August	106.6	112.2	104.4	109.2	105.5	101.8
September	107.0	112.4	104.7	108.4	105.0	102.8
October	108.1	113.0	104.6	110.6	105.1	101.5
November	108.0	113.0	105.2	111.6	103.8	103.6
December	108.0	112.5	105.4	111.1	103.7	104.0

Data table for Chart 2

Data table for chart 2
Table summary
This table displays the results of Data table for chart 2 Employees aged 15 to 29, Employees aged 30 to 49, High exposure and high complementarity, High exposure and low complementarity, Low exposure, High exposure and high complementarity, High exposure and low complementarity and Low exposure, calculated using index units of measure (appearing as column headers).

	Employees aged 15 to 29			Employees aged 30 to 49
	High exposure and high complementarity	High exposure and low complementarity	Low exposure	High exposure and high complementarity	High exposure and low complementarity	Low exposure
	index
Notes: The dashed vertical line demarcates the beginning of mass availability of ChatGPT in November 2022. The solid horizontal line demarcates the baseline of the employment growth index. Any point above the line indicates a higher employment level relative to the one observed in November 2022 and any point below the line indicates a lower employment level. Employment refers to the main job of employees. Employment growth is expressed relative to the level observed in November 2022. For example, an index of 110 would indicate 10% growth (110 minus 100) in employment relative to the November 2022 level, and an index of 90 would indicate that employment was 10% lower (90 minus 100) relative to the November 2022 level. Employment was disaggregated based on the complementarity-adjusted artificial intelligence occupational exposure (C-AIOE) index developed by Felten et al. (2021) and Pizzinelli et al. (2023). An occupation has potentially high exposure if its AIOE index exceeds the median AIOE across all occupations; otherwise, it is considered to have low exposure. Similarly, an occupation is considered to have potentially high complementarity if its complementarity parameter exceeds the median complementarity across all occupations; otherwise, it is considered to have low complementarity. Sources: Statistics Canada, Labour Force Survey, January 2015 to December 2025; and Occupational Information Network.
2015
January	80.1	82.2	97.2	76.2	83.2	99.7
February	81.5	82.8	97.0	76.1	84.6	98.8
March	82.1	82.2	96.8	76.4	82.6	99.2
April	82.0	81.0	97.4	76.6	83.2	98.5
May	82.3	82.5	97.1	77.4	82.7	98.0
June	84.1	82.0	96.7	78.2	82.7	97.8
July	88.5	82.1	96.6	77.9	84.3	97.3
August	88.8	82.8	96.4	79.1	82.2	97.3
September	86.0	82.9	96.1	79.8	82.9	96.9
October	85.8	83.2	95.2	79.7	83.1	96.6
November	86.3	81.6	94.6	79.6	82.8	97.2
December	83.8	82.5	94.8	79.8	83.6	96.6
2016
January	83.3	79.2	97.0	79.5	83.2	96.1
February	80.5	83.7	95.4	79.3	83.6	95.8
March	81.9	85.1	95.0	79.5	83.3	95.4
April	83.8	85.3	95.4	79.2	83.0	95.6
May	86.8	84.9	94.3	79.5	83.7	96.0
June	84.0	84.5	95.0	79.6	81.7	97.2
July	83.8	82.8	95.6	80.0	83.7	96.3
August	82.4	84.4	96.2	80.1	84.5	95.5
September	83.2	85.6	96.5	79.3	83.7	97.0
October	83.3	86.0	96.1	79.4	82.3	98.6
November	82.0	87.1	96.9	78.2	85.0	97.3
December	83.4	88.0	96.2	79.3	84.7	97.1
2017
January	84.3	87.5	96.9	79.8	85.4	97.8
February	84.6	86.8	97.2	80.6	84.8	98.2
March	85.6	85.9	97.8	80.8	85.9	98.1
April	85.0	86.5	97.3	81.0	85.5	98.6
May	85.2	86.8	99.1	80.7	85.2	99.1
June	84.2	87.7	98.8	79.6	85.4	99.1
July	83.9	88.8	98.7	81.1	85.0	99.1
August	84.6	89.0	97.7	80.3	86.0	97.4
September	84.2	89.1	98.3	80.1	85.7	97.4
October	85.3	88.8	99.2	80.4	85.8	97.7
November	87.7	85.9	101.7	81.9	85.2	97.2
December	88.5	88.3	99.9	82.5	84.6	98.6
2018
January	88.2	89.7	98.5	81.0	84.3	98.6
February	87.7	88.7	98.7	80.5	84.3	99.0
March	89.9	89.7	98.3	80.3	85.0	100.3
April	88.2	90.4	98.9	80.8	85.7	98.9
May	87.6	92.0	99.1	82.7	85.7	98.9
June	88.4	93.0	99.9	81.0	86.9	99.3
July	90.7	92.4	102.7	80.4	88.2	99.6
August	87.8	92.3	101.1	80.4	87.7	101.0
September	85.9	90.5	101.7	81.0	88.0	101.2
October	86.9	88.1	101.9	80.8	87.2	101.6
November	85.7	91.3	101.6	80.8	87.2	102.5
December	86.4	89.5	101.5	81.6	87.8	101.9
2019
January	87.8	91.8	101.2	83.6	87.5	101.9
February	88.7	94.0	100.4	83.2	88.0	100.9
March	86.2	94.6	99.2	83.2	87.2	102.1
April	91.0	95.7	101.3	83.9	87.5	103.0
May	93.8	93.7	102.3	83.4	89.4	102.6
June	95.0	93.5	101.2	84.3	88.8	102.4
July	95.5	94.4	100.1	84.5	87.8	103.1
August	95.4	95.1	100.0	84.2	87.5	103.1
September	96.8	93.3	98.7	84.6	87.4	102.9
October	94.0	93.4	98.4	85.5	88.1	101.4
November	94.2	94.5	98.3	85.8	88.3	101.1
December	92.9	97.3	98.8	85.7	89.7	100.9
2020
January	92.0	93.9	100.0	85.2	91.0	101.1
February	93.7	93.7	100.2	83.9	91.3	101.4
March	84.7	85.9	83.8	81.9	91.4	92.8
April	78.1	70.3	67.3	76.8	85.1	77.2
May	78.8	70.0	68.8	78.4	86.2	80.6
June	84.4	79.5	81.1	79.8	90.6	90.1
July	85.4	84.4	85.0	84.6	90.3	90.3
August	85.7	87.8	86.9	85.0	92.4	91.5
September	88.1	90.9	91.3	85.8	96.0	92.7
October	88.9	91.8	89.3	87.0	95.0	94.5
November	89.0	91.4	89.2	85.5	95.8	95.1
December	89.1	92.4	86.3	88.7	94.9	94.5
2021
January	87.8	87.1	84.9	86.2	95.9	93.9
February	86.4	89.6	89.2	87.5	97.2	92.8
March	87.9	96.7	89.3	88.5	97.8	94.5
April	92.8	91.8	89.0	88.5	98.3	92.4
May	88.4	92.6	88.2	87.4	99.5	91.4
June	91.1	95.5	94.5	88.4	98.3	94.8
July	92.2	97.1	94.8	89.8	98.6	94.2
August	95.2	98.3	94.8	90.5	96.9	94.2
September	96.0	100.8	94.7	90.6	98.7	94.6
October	90.5	101.8	95.0	91.3	99.6	94.5
November	92.2	100.2	98.4	91.9	100.4	94.9
December	100.4	99.6	97.0	92.7	100.4	98.0
2022
January	99.6	99.8	90.7	92.2	99.9	96.7
February	98.7	100.7	97.6	94.1	99.3	98.5
March	100.6	98.6	99.0	95.2	99.5	98.2
April	101.9	98.4	99.8	95.7	100.0	99.4
May	107.9	101.4	98.6	96.6	99.6	99.5
June	104.4	99.9	99.4	97.5	102.2	98.5
July	101.9	98.9	100.2	97.5	102.2	99.2
August	100.1	101.3	98.6	98.6	102.1	97.2
September	95.6	95.5	99.8	99.0	102.0	98.9
October	100.5	98.2	99.1	101.5	100.2	99.9
November	100.0	100.0	100.0	100.0	100.0	100.0
December	99.8	100.3	102.6	99.3	101.3	100.4
2023
January	99.7	102.2	101.5	99.2	102.7	104.1
February	101.6	102.9	100.8	99.8	99.5	102.5
March	99.3	104.4	100.9	100.1	101.3	102.0
April	99.3	104.1	100.5	100.6	102.3	101.2
May	101.5	100.1	102.1	102.3	102.7	101.6
June	101.4	103.5	101.1	103.2	102.1	101.5
July	104.1	102.7	101.7	103.6	102.8	101.9
August	104.3	100.2	104.0	103.7	103.6	102.0
September	105.2	99.0	103.7	104.4	103.5	102.7
October	106.4	98.2	104.1	104.3	104.8	102.7
November	106.2	99.1	103.8	104.8	105.1	103.3
December	107.2	100.3	103.8	106.0	103.9	105.1
2024
January	108.2	100.9	102.5	107.2	103.6	103.3
February	107.3	100.8	103.7	107.4	104.4	100.7
March	109.8	100.2	103.2	108.1	104.7	102.7
April	110.0	101.0	104.4	108.4	104.6	104.3
May	108.5	100.7	105.5	107.7	103.3	104.4
June	104.3	99.7	106.3	107.2	103.7	105.4
July	109.5	101.9	105.4	107.1	104.7	105.5
August	107.9	101.7	105.3	107.4	105.1	106.7
September	106.7	103.6	105.1	106.9	105.8	106.3
October	105.3	104.4	105.2	107.1	105.0	106.5
November	107.9	103.8	104.5	107.0	107.0	106.9
December	105.9	107.4	104.5	107.3	108.4	107.4
2025
January	105.9	103.0	106.0	107.7	109.6	108.3
February	111.2	103.1	105.7	107.3	109.1	109.0
March	107.2	103.1	105.9	106.6	109.9	110.3
April	106.1	104.8	105.7	106.7	109.3	109.9
May	106.3	104.7	105.0	107.1	111.7	108.0
June	106.3	104.7	105.4	107.8	112.6	109.2
July	106.1	102.4	105.7	108.3	111.5	108.9
August	108.8	103.8	102.8	108.5	112.3	107.8
September	108.3	102.7	105.0	109.3	113.1	107.9
October	111.5	102.9	103.4	109.9	113.2	107.8
November	111.3	102.4	107.7	111.0	112.2	107.8
December	112.5	100.0	108.0	110.2	112.9	107.2

Data table for Chart 3

Data table for chart 3
Table summary
This table displays the results of Data table for chart 3 Employees whose highest level of education was a bachelor's degree or above, Employees whose highest level of education was a postsecondary credential below a bachelor's degree or some postsecondary education (incomplete), Employees whose highest level of education was a trades certificate or diploma, Employees whose highest level of education was a high school diploma or below, High exposure and high complementarity, High exposure and low complementarity, Low exposure, High exposure and high complementarity, High exposure and low complementarity, Low exposure, High exposure and high complementarity, High exposure and low complementarity, Low exposure, High exposure and high complementarity, High exposure and low complementarity and Low exposure, calculated using index units of measure (appearing as column headers).

	Employees whose highest level of education was a bachelor's degree or above			Employees whose highest level of education was a postsecondary credential below a bachelor's degree or some postsecondary education (incomplete)			Employees whose highest level of education was a trades certificate or diploma			Employees whose highest level of education was a high school diploma or below
	High exposure and high complementarity	High exposure and low complementarity	Low exposure	High exposure and high complementarity	High exposure and low complementarity	Low exposure	High exposure and high complementarity	High exposure and low complementarity	Low exposure	High exposure and high complementarity	High exposure and low complementarity	Low exposure
	index
Notes: The dashed vertical line demarcates the beginning of mass availability of ChatGPT. The solid horizontal line demarcates the baseline of the employment growth index. Any point above the line indicates a higher employment level relative to the one observed in November 2022 and any point below the line indicates a lower employment level. Employment refers to the main job of employees aged 15 and older. Employment growth is expressed relative to the level observed in November 2022. For example, an index of 110 would indicate 10% growth (110 minus 100) in employment relative to the November 2022 level, and an index of 90 would indicate that employment was 10% lower (90 minus 100) relative to the November 2022 level. Employment was disaggregated based on the complementarity-adjusted artificial intelligence occupational exposure (C-AIOE) index developed by Felten et al. (2021) and Pizzinelli et al. (2023). An occupation has potentially high exposure if its AIOE index exceeds the median AIOE across all occupations; otherwise it is considered to have low exposure. Similarly, an occupation is considered to have potentially high complementarity if its complementarity parameter exceeds the median complementarity across all occupations; otherwise it is considered to have low complementarity. Sources: Statistics Canada, Labour Force Survey, January 2015 to December 2025; and Occupational Information Network.
2015
January	70.1	64.9	72.8	89.8	93.7	93.4	89.2	97.8	101.4	87.0	107.1	106.7
February	71.0	64.5	72.9	90.0	94.5	93.4	94.1	94.7	100.9	88.1	106.8	106.2
March	72.0	64.5	73.4	89.8	93.4	93.8	85.5	93.6	101.4	92.1	105.9	105.9
April	73.1	65.0	74.2	89.9	93.2	93.8	77.5	93.7	101.9	87.2	102.5	105.7
May	73.9	65.8	71.9	90.5	93.7	93.4	80.1	89.0	100.9	87.5	101.3	106.0
June	74.9	66.9	73.9	90.7	93.8	92.5	77.8	92.4	101.2	89.1	100.4	106.1
July	75.3	68.1	74.1	90.8	92.5	91.7	79.0	93.8	101.5	89.6	101.8	105.8
August	77.9	69.5	72.7	89.9	90.3	92.4	83.1	94.9	100.8	89.3	101.5	105.5
September	77.7	69.1	71.6	88.7	91.4	92.0	86.3	95.7	99.0	89.3	100.7	105.8
October	76.9	70.3	71.6	87.2	92.8	91.1	90.9	94.9	100.6	91.0	100.8	105.1
November	77.2	70.1	70.9	87.4	93.7	91.6	91.8	92.8	102.5	89.0	99.7	103.8
December	76.5	70.3	70.6	86.7	92.5	90.9	92.7	93.0	100.3	91.6	100.5	103.4
2016
January	76.5	70.7	72.1	86.7	92.4	94.7	90.5	91.0	98.5	89.2	98.4	103.6
February	75.9	71.7	72.2	89.3	92.7	93.6	88.6	92.1	98.1	86.5	97.7	104.3
March	76.2	72.6	72.3	87.4	94.0	93.7	87.7	90.4	96.9	85.5	97.7	103.2
April	76.9	72.5	70.4	87.9	94.1	93.3	87.3	85.9	97.9	85.0	102.7	103.3
May	79.1	72.1	73.0	86.5	94.3	92.9	87.7	90.3	98.2	82.6	101.6	103.3
June	77.8	71.9	71.3	88.8	93.8	93.3	89.9	90.0	97.8	81.7	101.1	103.6
July	77.6	71.7	73.9	88.9	92.7	93.2	89.0	90.5	97.9	82.8	100.4	103.8
August	77.7	71.1	74.5	89.0	96.2	92.6	87.1	90.0	98.6	85.7	102.8	103.8
September	76.7	70.8	76.5	89.7	95.0	93.5	84.9	90.2	100.7	83.9	104.1	103.4
October	76.8	72.3	75.1	90.2	93.9	94.9	80.4	90.4	100.5	84.1	102.5	103.9
November	75.8	76.2	74.8	89.0	93.7	94.6	75.2	92.7	100.0	85.4	103.1	104.5
December	76.6	74.3	74.9	90.1	93.9	94.2	77.5	93.5	98.7	84.2	102.9	105.4
2017
January	77.4	74.0	74.0	90.3	95.5	94.7	80.8	92.5	102.1	84.4	104.3	105.7
February	77.8	72.8	74.7	90.2	93.7	93.7	85.8	90.3	104.5	86.0	104.0	106.3
March	78.2	74.4	72.8	89.6	93.4	93.5	81.6	93.7	104.1	86.8	102.8	107.4
April	77.8	73.0	74.4	89.5	93.4	94.0	84.4	92.8	103.0	88.0	102.6	107.7
May	77.7	75.3	76.6	89.8	92.3	95.6	90.5	94.6	104.0	89.5	102.3	107.0
June	77.0	77.3	76.3	89.2	93.9	95.9	85.1	94.9	103.9	89.7	102.4	106.4
July	79.0	75.1	76.8	89.2	94.8	96.3	84.4	94.9	102.3	88.8	102.2	106.5
August	78.4	75.7	76.9	89.1	95.1	97.1	86.1	94.7	101.5	93.7	100.7	106.0
September	79.2	75.7	78.0	87.9	94.5	96.6	86.4	94.7	101.4	90.1	100.6	106.0
October	79.6	75.6	78.4	87.2	94.9	96.5	87.2	94.3	101.3	89.2	102.7	106.0
November	81.0	75.3	79.4	89.8	94.4	96.6	88.0	92.4	101.5	83.0	101.4	106.5
December	81.4	75.8	79.8	89.7	95.0	97.7	91.6	90.0	101.0	86.1	100.2	106.3
2018
January	80.6	76.2	79.9	90.4	95.2	97.8	95.3	91.9	101.7	88.4	101.3	104.7
February	80.6	76.2	80.1	90.7	94.6	98.4	90.0	88.3	102.2	87.7	102.5	103.6
March	80.4	76.8	80.0	91.6	94.0	98.9	89.5	90.0	105.2	87.5	104.0	103.0
April	80.3	77.9	80.2	92.0	94.6	100.3	89.0	90.4	103.2	86.5	103.9	102.6
May	82.2	77.2	79.2	90.4	95.4	99.8	86.7	88.8	103.6	88.7	104.1	103.9
June	80.1	78.7	77.1	90.5	95.6	100.7	85.0	95.2	104.3	88.5	105.4	104.1
July	79.1	79.2	81.2	89.6	96.7	101.7	85.4	89.0	106.4	88.3	106.0	104.1
August	79.2	79.3	83.2	88.3	97.1	101.5	81.9	89.3	105.8	87.5	102.6	104.7
September	79.2	79.4	84.3	89.7	96.8	101.8	80.3	88.0	104.1	85.8	100.2	105.3
October	79.5	78.8	85.3	89.8	96.1	104.6	86.4	84.5	103.7	86.2	99.9	104.3
November	79.3	78.7	87.3	90.0	97.6	103.7	87.1	89.1	104.1	87.0	101.4	103.3
December	80.5	78.2	87.8	90.5	95.1	102.6	86.6	89.1	104.2	88.7	101.5	104.4
2019
January	81.7	79.3	88.5	92.6	96.3	102.6	86.5	87.6	104.0	87.8	100.9	106.2
February	82.3	80.4	88.5	89.9	98.3	103.1	90.1	91.2	102.5	87.4	99.9	104.1
March	82.5	80.3	86.7	87.9	99.8	104.1	91.7	92.2	101.6	87.2	97.6	104.0
April	83.3	81.6	92.0	91.1	100.0	103.3	89.5	92.8	104.0	88.5	99.2	104.3
May	83.7	84.1	91.7	92.8	99.9	104.4	90.3	93.3	104.1	88.1	98.8	103.5
June	84.7	84.2	92.8	92.2	100.8	103.2	90.4	92.9	101.6	88.4	96.9	102.9
July	84.4	84.4	91.9	93.4	98.8	103.3	89.3	91.7	105.3	88.0	96.0	102.2
August	84.8	85.6	92.5	93.5	97.7	102.3	90.0	91.9	106.9	85.7	97.7	103.2
September	84.8	85.0	92.7	94.6	97.4	102.5	88.9	90.6	107.8	85.5	96.8	100.8
October	85.2	86.2	92.9	92.9	97.9	101.4	91.4	91.9	106.7	83.8	97.4	100.5
November	85.5	86.6	92.3	92.4	97.3	99.3	92.3	92.3	106.9	83.0	96.4	100.1
December	84.2	87.5	94.6	91.5	98.8	102.2	93.9	95.1	104.8	84.9	97.6	100.4
2020
January	84.5	87.5	96.4	90.2	99.2	102.3	92.7	99.3	103.8	87.3	98.1	101.6
February	84.6	87.2	96.7	90.8	97.5	101.6	90.8	94.4	102.8	85.3	102.1	102.5
March	81.5	87.4	83.6	86.0	91.9	90.6	90.3	91.1	96.9	80.9	93.8	90.8
April	79.9	83.0	73.5	79.1	81.1	76.1	79.8	73.3	76.0	73.1	79.0	73.3
May	82.4	84.4	75.0	80.1	83.3	76.7	78.6	84.2	85.3	76.5	77.8	76.0
June	83.9	85.2	82.5	82.5	92.9	82.7	95.2	88.1	95.2	78.1	88.4	86.7
July	86.9	84.7	84.4	85.3	92.8	86.3	97.2	93.1	97.8	75.4	94.5	87.4
August	87.4	88.2	85.1	85.7	94.6	89.6	102.1	91.9	98.5	77.5	95.2	88.7
September	88.2	91.3	86.3	85.7	95.9	92.4	97.3	95.8	97.8	79.9	99.5	92.6
October	88.8	93.4	86.0	85.9	96.0	93.1	94.5	91.7	99.2	80.9	96.3	91.5
November	89.0	95.1	87.2	85.3	96.8	93.0	92.6	89.8	98.2	80.2	97.1	91.9
December	91.7	96.4	86.3	85.1	95.4	93.2	88.4	88.0	98.7	78.4	94.8	91.2
2021
January	90.3	96.5	86.4	85.4	93.0	93.4	81.4	87.2	98.9	79.0	90.2	90.2
February	90.9	97.4	88.3	85.7	96.6	94.7	88.7	88.9	99.1	79.6	92.6	90.9
March	91.0	99.1	88.7	87.0	97.8	96.7	87.9	90.3	98.8	85.5	94.4	91.6
April	91.9	98.6	88.6	88.0	98.4	94.8	90.2	86.4	96.1	82.8	92.6	91.3
May	89.2	100.2	90.8	89.2	99.7	93.1	90.0	87.0	96.1	82.3	93.4	90.7
June	90.8	100.6	91.9	93.0	100.3	96.1	94.5	86.2	95.6	81.3	94.4	94.6
July	92.0	101.5	89.9	91.3	100.3	94.7	88.4	87.4	93.1	86.0	96.7	95.6
August	93.7	100.6	94.2	93.5	100.2	95.2	82.3	86.9	93.1	88.7	98.6	95.1
September	93.3	100.9	94.0	93.8	103.1	94.0	89.9	90.7	93.9	92.0	101.4	95.5
October	91.8	100.5	96.0	94.8	101.9	94.5	88.1	90.3	94.2	93.7	101.0	96.4
November	93.3	100.3	96.0	96.4	100.9	96.6	88.6	91.4	95.7	94.5	102.3	97.3
December	94.5	99.8	102.1	97.4	100.9	96.5	92.6	92.0	95.6	94.0	104.5	97.7
2022
January	94.9	99.8	90.4	98.2	101.1	94.9	93.1	92.6	95.6	93.2	102.1	96.3
February	95.0	101.6	98.2	98.7	100.5	98.0	94.8	91.8	97.0	94.8	103.0	99.2
March	96.7	99.7	100.1	98.1	100.2	98.1	96.0	92.2	97.7	92.6	102.0	100.1
April	97.4	100.6	100.4	100.1	99.2	99.2	96.5	94.0	98.7	92.6	104.6	100.0
May	99.0	100.3	100.0	98.5	102.4	98.3	105.0	92.6	99.9	94.1	107.4	99.5
June	100.8	100.3	100.0	98.6	103.7	97.8	97.8	94.8	99.3	94.5	103.0	97.7
July	99.2	100.2	98.7	95.2	102.6	99.4	99.0	94.4	100.2	94.0	102.4	99.1
August	99.4	102.3	97.7	98.4	101.6	97.4	99.9	99.3	99.3	95.7	101.4	100.1
September	99.0	100.5	98.1	99.1	100.7	99.0	95.0	97.1	100.6	94.6	99.2	100.6
October	99.8	100.6	98.7	100.1	98.3	99.6	99.7	99.2	99.9	96.3	100.3	100.3
November	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
December	99.2	101.9	100.7	99.8	99.9	101.0	100.0	108.0	101.0	96.8	99.1	101.4
2023
January	99.5	102.7	102.5	100.1	100.3	102.6	103.0	103.4	102.3	95.7	103.1	102.9
February	100.1	101.8	108.2	101.0	101.2	103.0	103.8	98.9	100.2	94.1	104.1	101.3
March	100.1	104.2	105.7	102.9	101.5	102.1	109.0	97.9	101.2	94.9	103.8	101.1
April	100.8	106.3	106.3	103.7	101.5	101.9	108.9	98.1	101.1	95.9	103.5	98.6
May	102.7	104.3	105.8	105.1	101.5	102.6	103.0	98.9	100.5	96.0	106.1	100.6
June	103.5	106.5	102.5	105.7	102.1	102.5	102.3	97.6	100.3	96.2	105.6	100.3
July	103.0	104.0	108.1	107.0	103.1	99.6	103.6	97.7	101.5	96.6	104.4	101.4
August	104.2	103.9	107.9	105.1	104.0	102.3	102.2	97.6	103.4	95.4	103.0	101.4
September	105.0	103.0	110.9	106.2	103.7	102.3	106.4	97.4	102.0	93.2	102.5	100.7
October	105.7	103.0	111.9	103.2	103.5	102.1	106.0	97.0	101.5	89.9	102.3	101.5
November	106.5	104.4	113.1	105.9	104.1	101.8	102.1	98.1	100.9	91.6	101.0	101.8
December	107.5	103.6	118.2	106.5	104.1	101.2	103.1	95.0	101.7	93.1	100.5	100.9
2024
January	108.9	105.2	116.3	107.7	103.3	100.2	104.1	94.5	102.0	95.3	100.5	101.1
February	109.9	106.2	116.1	108.3	103.2	99.3	101.6	96.8	103.4	93.8	100.7	99.9
March	110.9	106.1	115.0	107.4	102.7	100.3	102.6	95.0	102.4	94.5	101.1	99.5
April	111.0	108.0	119.9	106.4	103.7	101.1	100.7	93.5	104.0	94.4	100.6	99.4
May	111.1	108.6	125.5	105.9	100.2	103.4	101.0	95.1	103.7	97.8	98.2	98.3
June	111.0	110.1	124.5	103.1	100.0	104.5	103.7	92.8	103.3	92.6	98.0	98.0
July	111.7	110.6	126.3	104.4	98.0	105.7	103.3	92.6	100.9	88.8	98.1	97.2
August	111.8	111.3	129.8	103.6	100.0	106.8	104.6	90.7	102.8	91.6	97.4	96.5
September	113.1	112.6	127.0	102.1	101.3	106.7	98.0	95.1	103.1	87.5	99.8	95.8
October	110.6	112.7	127.3	101.7	101.8	107.0	103.7	90.3	105.1	91.0	99.0	94.4
November	111.3	113.1	127.9	101.5	102.9	105.5	106.8	89.1	104.8	93.1	99.3	95.8
December	111.5	114.2	125.3	102.5	103.3	104.6	106.1	96.8	104.8	90.5	101.6	96.0
2025
January	110.7	114.8	125.4	101.8	102.8	105.9	103.4	94.0	106.0	94.2	101.3	96.9
February	110.5	114.7	130.6	101.3	101.6	104.9	107.5	94.8	105.9	99.5	101.0	97.5
March	109.8	115.7	125.1	100.2	100.5	104.5	101.5	98.5	112.8	97.7	99.1	97.5
April	109.8	114.9	124.9	97.5	102.1	103.8	108.7	102.1	107.8	98.2	102.9	98.9
May	109.7	116.9	123.9	99.0	102.2	100.4	113.8	100.7	109.5	98.9	104.5	99.9
June	110.3	115.6	124.4	99.2	103.9	99.0	115.4	103.5	111.5	100.0	105.2	100.4
July	110.8	114.6	124.6	98.9	101.3	97.3	114.5	102.3	112.4	101.6	105.0	100.0
August	111.3	117.7	122.7	99.4	100.6	95.4	117.4	102.8	112.3	102.6	105.6	100.5
September	111.7	119.6	124.7	98.1	99.5	95.8	119.2	100.3	111.3	103.9	104.2	101.2
October	112.7	119.3	124.9	100.4	99.7	95.1	117.1	107.0	110.8	105.0	103.8	100.5
November	113.9	118.0	125.0	101.7	98.8	99.1	120.9	102.8	113.7	103.4	103.9	100.0
December	114.1	118.5	126.6	98.9	98.8	97.5	119.2	102.5	114.6	104.2	101.9	100.5

Employment in larger establishments grew relatively faster from November 2022 to December 2025 regardless of potential occupational exposure to and complementarity with artificial intelligence

There is some evidence that larger establishments may be more likely to adopt AI than their smaller counterparts (Statistics Canada, 2025a). Chart 4 examines the relationship between potential occupational exposure to and complementarity with AI by establishment size, distinguishing between employees in establishments with 500 or fewer workers and those in establishments with more than 500 workers. Public administration, educational services, and health care and social assistance were excluded in this part of the analysis to reduce the potential influence of public sector employers which tend to be larger.

From November 2022 to December 2025, there was little to no employment growth in smaller establishments whereas employment grew by roughly 30% in larger establishments, regardless of potential occupational exposure to and complementarity with AI. These divergent growth rates reflect longer-term trends observed over the past decade. In particular, HELC and HEHC jobs grew significantly faster in larger establishments with the number of jobs in each of these occupational groups more than doubling over the past decade.

Employment growth in industries potentially more exposed to and less complementary with artificial intelligence was not significantly different than that in other industries from the fourth quarter of 2022 to the third quarter of 2025

Firm-level AI adoption rates vary considerably across industries (Statistics Canada, 2024a; 2024b). To get a sense of whether the advent of generative AI has slowed employment and earnings growth in industries potentially more exposed to and less complementary with AI, the following linear regression model was estimated at the industry level, as measured by the four-digit North American Industry Classification System excluding public administration, educational services, and health care and social assistance:

$$\Delta \ln y_{2022q4to2025q3,i}=\alpha +{\beta }_{1}HEL{C}_{2021,i}+{\beta }_{2}HEH{C}_{2021,i}+\gamma \Delta \ln y_{2016q4to2019q3,i}+\theta X+u_i$$

where $\Delta \ln y_{2022q4to2025q3,i}$ is the change in the natural logarithm of a given outcome in industry $i$ from the fourth quarter of 2022 to the third quarter of 2025.^Note  Two outcomes were considered: (1) employment and (2) average weekly earnings. HELC2021, $i$ and HEHC2021,$i$ denote the share of HELC and HEHC jobs, respectively, in industry $i$ from 2021 (the share of LE jobs was excluded to avoid multicollinearity). The term $\gamma \Delta \ln y_{2016q4to2019q3,i}$ denotes the change in the log outcome from the fourth quarter of 2016 to the third quarter of 2019 for capturing pre-trends before the pandemic. The last term, $X$ , captures within industry $i$ the share of teleworkable jobs in 2021 based on telework potential (Dingel and Neiman, 2020), the share of unionized jobs in 2021, and the share of employees in enterprises with more than 500 workers in 2022.

The outcome and pre-trend data were obtained from the Survey of Employment, Payrolls and Hours (Table 14-10-0220-01). The share of employees in enterprises with more than 500 workers was obtained from the 2022 Longitudinal Worker File (the latest available year at the time). The remaining covariates were obtained from the 2021 Census of Population. The share of HELC and HEHC occupations within industries remained stable across the 2016 and 2021 censuses of population, so it is unlikely that these would have changed in any significant way from 2021 to 2025.

Accounting for pre-trends in employment growth before the pandemic, firm size, unionization and the potential for telework across industries, a one percentage point increase in the share of HELC jobs was associated with a 0.039% decrease in employment growth from the fourth quarter of 2022 to the third quarter of 2025. By contrast, a one percentage point increase in the share of HEHC jobs was associated with a 0.104% increase in employment growth over the same period. However, neither of these results was statistically significant. Accounting for pre-trends in average weekly earnings growth before the pandemic, firm size, unionization and the potential for telework across industries, a one percentage point increase in the share of HELC jobs was associated with a 0.038% decrease in average weekly earnings growth, while a one percentage point increase in the share of HEHC jobs was associated with a 0.057% decrease. Once again, neither effect was statistically significant. These results suggest that there is no clear evidence of a slowdown in employment or weekly earnings growth in industries potentially more exposed to and less complementary with AI.

Coding-intensive jobs grew at a similar rate as other jobs from November 2022 to December 2025

Given AI’s growing coding and debugging capabilities, there are concerns that AI could be supplanting the role of computer programmers and other coding-intensive jobs that are a subset of HELC jobs described in footnote 5. From November 2022 to December 2025, coding-intensive professions grew by roughly 15% compared with a growth rate of about 5% for other jobs, albeit the difference was not statistically significant (Chart 5). Over the longer term, coding-intensive professions grew substantially faster than other jobs over the last decade. And coding-intensive jobs were largely shielded from the immediate negative shocks of the pandemic relative to other jobs.

There are growing concerns regarding AI replacing entry-level tasks such as writing basic code. Brynjolfsson et al. (2025) found that the relative number of software developers younger than 30 in the United States has decreased since November 2022. Chart 6 shows job growth in Canadian coding-intensive professions by age group. Employees aged 30 to 49 generally saw more robust employment gains compared with their younger counterparts aged 15 to 29. In December 2025, employment among those aged 15 to 29 in coding-intensive jobs was about the same as it was in November 2022, while employment among those aged 30 to 49 was almost 30% higher. This divergence has been statistically significant since late 2024. In absolute terms, there were twice as many employees aged 30 to 49 in coding-intensive jobs in November 2022 relative to those aged 15 to 29. By December 2025, there were three times as many. For other jobs, this ratio remained stable at around 1.5 over the last decade.

Data table for Chart 4

Data table for chart 4
Table summary
This table displays the results of Data table for chart 4 Establishments with 500 or fewer workers, Establishments with more than 500 workers, High exposure and high complementarity, High exposure and low complementarity, Low exposure, High exposure and high complementarity, High exposure and low complementarity and Low exposure, calculated using index units of measure (appearing as column headers).

	Establishments with 500 or fewer workers			Establishments with more than 500 workers
	High exposure and high complementarity	High exposure and low complementarity	Low exposure	High exposure and high complementarity	High exposure and low complementarity	Low exposure
	index
Notes: The dashed vertical line demarcates the beginning of mass availability of ChatGPT in November 2022. The solid horizontal line demarcates the baseline of the employment growth index. Any point above the line indicates a higher employment level relative to the one observed in November 2022 and any point below the line indicates a lower employment level. Employment refers to the main job of employees aged 15 and older. Public administration, educational services, and health care and social assistance were excluded since the majority of employers in these industries employ more than 500 workers which would make comaring panel (a) and (b) challenging. Employment growth is expressed relative to the level observed in November 2022. For example, an index of 110 would indicate 10% growth (110 minus 100) in employment relative to the November 2022 level, and an index of 90 would indicate that employment was 10% lower (90 minus 100) relative to the November 2022 level. Employment was disaggregated based on the complementarity-adjusted artificial intelligence occupational exposure (C-AIOE) index developed by Felten et al. (2021) and Pizzinelli et al. (2023). An occupation has potentially high exposure if its AIOE index exceeds the median AIOE across all occupations; otherwise it is considered to have low exposure. Similarly, an occupation is considered to have potentially high complementarity if its complementarity parameter exceeds the median complementarity across all occupations; otherwise it is considered to have low complementarity. Sources: Statistics Canada, Labour Force Survey, January 2015 to December 2025; and Occupational Information Network.
2015
January	80.0	87.9	101.8	51.6	62.3	87.8
February	80.7	88.4	101.3	54.0	64.2	89.1
March	81.3	87.5	101.6	55.7	64.8	89.4
April	81.3	86.8	101.5	55.8	63.1	91.9
May	81.2	87.0	101.0	59.0	64.7	92.9
June	84.2	87.1	101.1	61.5	63.8	90.7
July	83.9	88.4	100.1	62.7	64.7	97.2
August	84.7	88.0	99.9	64.7	64.5	96.4
September	84.0	89.2	99.0	63.3	63.1	97.2
October	84.1	89.1	98.8	61.1	65.4	95.4
November	85.3	88.7	98.4	60.7	67.7	97.5
December	84.6	89.3	97.7	60.1	66.2	96.0
2016
January	82.7	87.8	99.6	62.5	66.9	89.2
February	82.0	89.0	99.9	63.3	68.5	94.0
March	81.3	89.7	98.9	64.4	70.4	95.0
April	81.7	90.5	98.9	67.2	65.8	93.7
May	81.9	90.9	99.1	65.1	63.8	94.5
June	82.6	90.6	99.4	66.7	64.5	94.0
July	82.9	90.5	99.7	66.5	65.9	89.7
August	82.5	91.2	100.0	67.8	66.7	89.4
September	80.9	91.5	100.5	66.0	64.8	90.2
October	80.3	91.5	100.6	68.6	62.9	93.0
November	80.6	92.6	100.7	69.1	66.9	90.6
December	81.3	92.3	100.6	72.2	69.3	91.3
2017
January	81.7	92.6	100.7	73.8	68.5	94.2
February	83.1	91.8	101.4	72.4	70.7	94.2
March	84.2	91.9	101.6	70.7	72.3	94.6
April	84.8	90.0	102.1	66.4	72.8	94.6
May	85.5	91.3	102.4	64.4	69.0	96.0
June	85.4	92.1	102.3	65.0	69.2	95.6
July	86.8	92.3	102.5	65.0	69.8	97.2
August	86.3	92.8	102.1	68.0	67.4	98.7
September	82.9	92.5	102.6	71.3	67.9	96.0
October	83.4	93.3	102.5	72.0	66.4	95.9
November	86.0	92.8	103.0	72.7	64.5	98.9
December	87.1	93.2	102.8	72.3	69.4	98.5
2018
January	87.2	93.9	101.9	69.2	68.8	98.8
February	85.8	94.4	101.9	72.6	68.3	94.3
March	85.7	94.5	102.2	73.2	67.5	97.8
April	84.9	95.1	102.0	72.5	66.2	100.2
May	85.3	95.1	102.7	72.5	66.7	98.9
June	84.2	95.4	103.2	74.1	70.6	101.5
July	82.6	95.0	103.9	75.0	75.9	101.6
August	82.0	94.9	104.0	71.6	71.1	100.5
September	83.6	94.1	104.2	68.5	71.3	103.7
October	83.9	92.1	104.6	71.1	77.1	102.1
November	84.6	94.1	104.9	71.7	73.8	99.6
December	85.5	92.3	105.1	70.8	76.5	101.8
2019
January	85.6	93.7	105.5	71.4	72.5	98.3
February	86.0	94.8	104.0	70.7	73.9	105.0
March	85.8	95.1	104.2	71.4	71.5	107.0
April	87.0	95.5	105.0	70.6	73.8	108.9
May	87.5	95.7	105.2	78.4	76.5	109.9
June	89.2	95.6	103.9	72.3	78.1	114.7
July	90.4	94.8	103.6	74.1	78.2	111.6
August	89.4	94.9	104.0	75.0	81.6	104.3
September	88.7	94.4	103.5	76.2	79.9	108.6
October	87.8	95.6	102.9	74.4	77.1	109.2
November	87.7	95.8	102.3	72.3	78.5	109.7
December	87.2	97.0	103.0	72.7	78.9	111.4
2020
January	87.3	97.7	103.2	68.3	76.7	105.0
February	86.9	97.2	103.8	72.1	82.2	101.8
March	85.3	93.6	91.8	73.7	80.6	96.8
April	78.5	79.6	72.7	67.5	81.4	78.3
May	78.4	82.3	76.9	70.7	80.5	83.3
June	82.6	90.7	86.9	73.6	76.6	89.0
July	84.7	91.9	89.7	79.5	80.1	88.6
August	86.5	94.7	91.5	76.5	80.2	89.2
September	87.2	97.6	94.7	68.6	83.2	90.5
October	87.6	98.0	94.2	73.7	81.2	88.0
November	86.9	98.6	94.7	74.1	84.9	85.3
December	87.2	99.4	93.5	76.5	79.2	85.8
2021
January	85.2	97.1	91.4	77.7	79.9	86.1
February	87.2	97.8	93.5	75.2	84.1	88.4
March	86.5	100.1	94.9	74.5	83.7	90.0
April	88.1	97.3	93.4	78.0	88.5	88.8
May	86.8	97.2	93.0	78.8	97.0	85.5
June	90.0	97.7	97.4	78.4	91.8	87.4
July	91.1	98.7	96.4	81.6	97.8	89.1
August	93.8	98.9	96.3	81.5	95.3	91.7
September	94.8	100.2	96.3	81.6	96.9	90.3
October	94.4	100.3	96.8	81.8	96.3	91.7
November	95.8	100.0	97.4	86.4	98.5	95.5
December	95.3	100.8	98.3	90.0	96.0	94.4
2022
January	95.3	100.3	94.6	87.3	99.5	92.5
February	95.3	100.3	99.4	90.1	105.3	94.3
March	96.8	99.0	99.9	86.3	102.2	97.1
April	95.8	101.2	100.5	91.4	100.6	98.0
May	99.0	102.5	100.0	94.6	99.0	101.0
June	99.0	101.3	99.8	92.3	102.6	104.8
July	99.5	101.2	100.9	92.4	98.8	99.2
August	99.3	102.1	100.0	95.0	102.7	92.9
September	99.4	99.8	100.8	92.0	95.0	96.4
October	100.8	99.8	100.5	94.6	98.0	98.3
November	100.0	100.0	100.0	100.0	100.0	100.0
December	98.0	99.8	102.5	98.6	104.6	99.0
2023
January	96.7	101.2	103.0	100.8	107.7	107.7
February	97.8	101.1	103.5	101.2	104.0	104.0
March	99.9	101.8	103.5	101.9	103.6	98.0
April	102.4	101.4	102.3	102.7	110.4	104.2
May	103.6	102.2	103.0	105.5	112.1	105.6
June	104.7	103.4	102.5	107.5	111.2	108.0
July	104.6	102.6	102.8	104.1	112.3	117.3
August	103.4	101.9	103.7	110.2	113.7	112.2
September	103.7	100.8	103.6	101.1	110.5	115.6
October	102.0	101.8	103.8	101.0	107.4	114.8
November	102.5	103.0	103.9	100.3	106.3	113.9
December	102.9	101.7	103.5	108.1	108.4	116.0
2024
January	103.9	101.5	103.3	108.3	109.1	114.2
February	106.2	102.0	103.2	108.5	107.5	109.6
March	106.0	101.8	103.0	112.2	103.2	107.5
April	105.9	102.3	104.7	108.8	106.8	108.2
May	106.9	101.4	105.5	112.7	104.2	106.1
June	104.8	102.2	105.9	113.2	100.2	105.7
July	105.6	101.4	105.9	108.5	107.4	105.4
August	104.0	102.4	105.8	113.2	105.9	107.5
September	102.4	103.3	105.1	110.6	107.8	113.3
October	101.8	103.1	105.5	113.1	106.7	106.3
November	102.7	103.6	104.7	111.2	111.8	105.2
December	100.6	104.7	104.7	116.0	117.2	107.1
2025
January	99.6	104.5	105.4	114.8	118.3	110.4
February	100.5	104.9	105.2	118.8	111.5	111.5
March	98.0	104.9	104.8	115.9	113.3	116.0
April	98.0	104.7	105.1	118.1	115.8	114.9
May	99.6	104.9	104.2	119.9	120.3	120.4
June	99.7	104.7	104.8	126.6	121.5	120.6
July	101.3	101.7	103.8	129.8	129.7	127.1
August	103.0	101.7	103.5	129.5	132.2	118.9
September	101.5	101.6	103.8	129.5	131.6	118.5
October	102.0	101.8	102.6	134.6	141.6	121.8
November	102.6	101.4	104.2	135.9	136.7	122.3
December	101.4	100.9	104.4	134.2	137.8	122.6

Data table for Chart 5

Data table for chart 5
Table summary
This table displays the results of Data table for chart 5 Other jobs and Coding-intensive jobs, calculated using index units of measure (appearing as column headers).

	Other jobs	Coding-intensive jobs
	index
Notes: The dashed vertical line demarcates the beginning of mass availability of ChatGPT in November 2022. The solid horizontal line demarcates the baseline of the employment growth index. Any point above the line indicates a higher employment level relative to the one observed in November 2022 and any point below the line indicates a lower employment level. Employment refers to the main job of employees aged 15 and older. The following 2021 National Occupational Classification occupations were selected as coding-intensive jobs: data scientists, cybersecurity specialists, business systems specialists, information systems specialists, database analysts and data administrators, computer systems developers and programmers, software engineers and designers, software developers and programmers, web designers, and web developers and programmers. Employment growth is expressed relative to the level observed in November 2022. For example, an index of 110 would indicate 10% growth (110 minus 100) in employment relative to the November 2022 level, and an index of 90 would indicate that employment was 10% lower (90 minus 100) relative to the November 2022 level. Employment was disaggregated based on the complementarity-adjusted artificial intelligence occupational exposure (C-AIOE) index developed by Felten et al. (2021) and Pizzinelli et al. (2023). An occupation has potentially high exposure if its AIOE index exceeds the median AIOE across all occupations; otherwise it is considered to have low exposure. Similarly, an occupation is considered to have potentially high complementarity if its complementarity parameter exceeds the median complementarity across all occupations; otherwise it is considered to have low complementarity. Sources: Statistics Canada, Labour Force Survey, January 2015 to December 2025; and Occupational Information Network.
2015
January	88.9	55.5
February	89.2	52.2
March	89.3	50.9
April	89.2	51.8
May	89.0	51.9
June	89.5	53.9
July	89.5	56.4
August	89.6	54.9
September	89.5	54.8
October	89.5	56.6
November	89.3	57.8
December	89.2	58.3
2016
January	89.2	57.4
February	89.2	62.1
March	89.2	60.0
April	89.5	58.8
May	89.7	56.2
June	89.7	55.0
July	89.7	57.1
August	90.0	55.5
September	90.2	54.5
October	90.4	54.3
November	90.5	58.5
December	90.8	59.3
2017
January	91.1	58.3
February	91.1	61.0
March	91.2	62.3
April	91.2	61.1
May	91.5	62.4
June	91.6	67.7
July	91.9	65.7
August	91.8	63.9
September	91.8	62.3
October	91.9	65.2
November	92.3	61.5
December	92.8	64.6
2018
January	92.4	64.5
February	92.4	63.2
March	92.7	64.6
April	92.7	62.7
May	92.7	63.5
June	93.1	62.8
July	93.5	65.6
August	93.3	64.4
September	93.6	59.8
October	93.8	61.6
November	94.0	61.9
December	94.1	63.4
2019
January	94.4	65.9
February	94.5	69.1
March	94.4	68.2
April	94.9	71.3
May	95.0	73.5
June	95.0	77.8
July	95.0	78.2
August	95.4	74.8
September	95.4	74.5
October	95.4	73.9
November	94.9	75.5
December	95.5	76.4
2020
January	95.3	78.1
February	95.4	80.4
March	88.6	82.0
April	77.1	82.4
May	79.0	82.2
June	85.6	86.1
July	87.9	85.0
August	89.6	85.4
September	92.0	89.7
October	92.2	91.5
November	92.4	91.5
December	92.2	93.5
2021
January	90.5	95.4
February	92.4	91.8
March	93.8	91.0
April	92.5	93.7
May	92.1	94.3
June	94.3	95.7
July	94.8	96.3
August	95.2	97.8
September	96.3	99.6
October	96.5	98.0
November	97.1	101.5
December	97.6	100.7
2022
January	95.9	103.6
February	98.1	100.9
March	98.3	100.9
April	98.7	100.8
May	99.0	105.2
June	98.9	110.9
July	99.2	106.4
August	99.3	105.1
September	99.4	103.5
October	99.9	101.4
November	100.0	100.0
December	100.2	102.2
2023
January	100.9	103.9
February	101.3	101.0
March	101.7	97.5
April	101.9	94.8
May	102.3	95.6
June	102.8	97.7
July	102.8	101.6
August	102.9	104.8
September	103.1	101.5
October	103.1	105.5
November	103.2	107.7
December	103.5	102.9
2024
January	103.7	102.8
February	103.8	103.8
March	104.0	101.3
April	104.5	103.6
May	104.3	104.9
June	104.4	105.8
July	104.5	105.6
August	105.1	102.9
September	105.0	109.1
October	104.9	109.6
November	104.9	110.6
December	105.3	108.9
2025
January	105.3	114.6
February	105.6	108.8
March	105.5	104.6
April	105.6	105.5
May	105.7	106.7
June	106.4	105.7
July	105.9	106.1
August	105.9	110.5
September	105.8	116.9
October	106.3	118.5
November	106.5	117.7
December	106.4	118.2

Data table for Chart 6

Data table for chart 6
Table summary
This table displays the results of Data table for chart 6 15 to 29, 30 to 49 and 50 and older, calculated using index units of measure (appearing as column headers).

	15 to 29	30 to 49	50 and older
	index
Notes: The dashed vertical line demarcates the beginning of mass availability of ChatGPT in November 2022. The solid horizontal line demarcates the baseline of the employment growth index. Any point above the line indicates a higher employment level relative to the one observed in November 2022 and any point below the line indicates a lower employment level. Employment refers to the main job of employees. The following 2021 National Occupational Classification occupations were selected as coding-intensive jobs: data scientists, cybersecurity specialists, business systems specialists, information systems specialists, database analysts and data administrators, computer systems developers and programmers, software engineers and designers, software developers and programmers, web designers, and web developers and programmers. Employment growth is expressed relative to the level observed in November 2022. For example, an index of 110 would indicate 10% growth (110 minus 100) in employment relative to the November 2022 level; an index of 90 would indicate that employment was 10% lower (90 minus 100) relative to the November 2022 level. Employment was disaggregated based on the complementarity-adjusted artificial intelligence occupational exposure (C-AIOE) index developed by Felten et al. (2021) and Pizzinelli et al. (2023). An occupation has potentially high exposure if its AIOE index exceeds the median AIOE across all occupations; otherwise it is considered to have low exposure. Similarly, an occupation is considered to have potentially high complementarity if its complementarity parameter exceeds the median complementarity across all occupations; otherwise it is considered to have low complementarity. Sources: Statistics Canada, Labour Force Survey, January 2015 to December 2025; and Occupational Information Network.
2015
January	43.0	61.5	52.3
February	37.5	58.1	52.2
March	40.1	57.4	45.6
April	39.8	58.0	50.2
May	41.4	58.3	50.9
June	44.5	59.0	51.3
July	43.8	63.3	51.8
August	49.7	58.2	54.6
September	48.4	58.5	54.9
October	50.1	62.6	50.5
November	53.5	63.2	51.9
December	54.2	62.9	49.8
2016
January	49.8	63.2	50.8
February	55.4	68.0	51.5
March	45.1	68.9	52.6
April	42.3	68.1	55.2
May	41.2	64.1	55.5
June	41.4	60.3	56.1
July	41.0	63.7	58.7
August	44.9	60.0	57.7
September	45.1	57.7	59.6
October	42.6	56.7	63.6
November	48.2	59.0	71.4
December	57.8	57.2	65.0
2017
January	57.9	56.5	61.6
February	60.2	59.9	63.8
March	60.2	61.2	65.7
April	52.6	64.2	64.5
May	59.0	64.3	62.7
June	58.9	71.1	68.2
July	62.3	68.5	63.1
August	64.6	64.3	63.8
September	61.8	63.6	62.5
October	69.3	64.9	63.7
November	60.0	63.8	61.3
December	61.9	65.9	63.8
2018
January	66.2	62.2	66.0
February	59.6	63.7	64.4
March	61.0	66.4	63.1
April	64.3	63.5	60.5
May	66.6	62.7	62.5
June	63.4	65.1	56.0
July	66.6	67.0	61.8
August	65.6	65.6	62.6
September	54.0	61.6	65.1
October	48.7	64.4	71.0
November	54.8	65.6	64.4
December	56.2	64.6	67.2
2019
January	55.3	70.2	63.9
February	62.6	72.5	65.3
March	67.5	69.3	66.0
April	70.6	72.5	69.5
May	68.2	76.5	71.3
June	78.6	80.0	72.0
July	84.2	78.4	72.6
August	76.2	77.1	71.7
September	66.9	81.5	68.7
October	73.7	78.9	65.1
November	80.1	79.1	65.6
December	80.2	78.9	65.5
2020
January	76.7	80.4	70.4
February	73.7	83.9	74.9
March	78.0	83.7	81.9
April	88.0	82.7	76.0
May	83.1	82.5	81.4
June	85.2	88.6	81.4
July	84.2	88.1	80.2
August	91.8	87.3	80.3
September	97.0	89.7	85.8
October	100.0	91.0	88.6
November	88.1	93.5	93.2
December	101.4	91.0	89.5
2021
January	92.5	99.1	86.0
February	77.1	100.4	83.7
March	86.9	95.8	82.5
April	95.7	98.3	79.7
May	102.2	97.5	79.0
June	102.5	99.6	80.2
July	93.3	105.0	81.0
August	91.0	107.1	88.5
September	95.8	103.2	99.7
October	90.0	105.6	93.1
November	85.4	110.0	98.0
December	89.1	107.1	95.9
2022
January	94.6	107.6	99.3
February	93.7	103.8	99.3
March	97.5	102.1	100.3
April	94.7	102.5	102.9
May	103.7	105.1	107.1
June	120.7	109.4	106.2
July	109.1	107.1	106.2
August	110.6	106.9	101.9
September	100.3	108.2	101.0
October	99.2	105.9	97.7
November	100.0	100.0	100.0
December	104.0	100.1	104.6
2023
January	101.4	103.1	103.4
February	102.4	95.8	110.7
March	98.8	94.7	101.0
April	88.7	95.4	100.2
May	94.9	95.9	94.4
June	98.4	100.0	93.6
July	101.0	106.6	94.7
August	101.6	112.9	95.8
September	92.2	111.8	92.2
October	96.5	115.0	95.3
November	90.9	120.4	94.0
December	86.7	115.5	89.8
2024
January	93.2	112.0	85.1
February	91.6	111.9	92.5
March	90.1	108.5	94.8
April	103.6	103.5	103.5
May	105.6	107.1	97.7
June	111.5	105.9	102.8
July	110.3	110.3	93.7
August	95.6	113.7	93.2
September	96.4	123.1	94.0
October	100.1	121.7	92.7
November	104.6	120.4	93.6
December	97.6	119.5	95.1
2025
January	94.2	129.3	95.7
February	90.2	121.3	95.9
March	80.0	118.2	96.6
April	83.0	119.1	95.5
May	80.0	121.1	99.1
June	87.6	116.7	100.0
July	80.9	120.9	101.1
August	91.9	126.0	102.1
September	105.3	127.5	106.8
October	106.1	128.0	110.4
November	104.5	128.0	107.2
December	108.4	127.0	106.3

However, it is not clear that the age divergence in coding-intensive professions is entirely caused by the advent of AI in 2022, because it was also around this time the economy was transitioning out of the pandemic and there was an increase in job vacancies, especially in technology-related jobs. It is possible that the relative decline in younger employees in coding-intensive jobs reflects a post-pandemic labour market adjustment process.

If AI replaces entry-level roles, this would introduce a new set of challenges regarding the development of senior-level expertise. If traditional entry-level positions are eliminated, new graduates could be expected to enter the workforce already equipped with advanced skills typically acquired through years of experience. This shift may disrupt traditional pathways for professional growth, necessitating new approaches to training and skill acquisition for future senior professionals.

Job vacancies in occupations potentially more exposed to and less complementary with artificial intelligence have decreased at a similar rate as those potentially less exposed from the fourth quarter of 2022 to the third quarter of 2025

Understanding labour market dynamics requires more than analyzing existing employment levels. Examining job vacancies offers a complementary perspective, particularly when evaluating the impact of emerging technologies such as AI on skills demand. Unlike employment, which reflects realized matches between workers and jobs, job vacancies provide a forward-looking measure of employers’ unmet demand for skills. Tracking vacancy trends can therefore shed light on whether demand for certain roles or skill sets is outpacing supply, highlight potential reallocation pressures within the labour market, and help identify occupations where technological disruption may be reshaping employer needs faster than workers can adapt.

The pandemic played a big role in restructuring skills demand over the last five years, as it may have accelerated technology adoption by businesses operating with reduced staff levels. Job vacancies started rising in late 2021 as many businesses started reopening. Vacancies reached a 10-year high of nearly 1 million by the second quarter of 2022 (Statistics Canada, 2025c). Chart 7 shows that this surge was largely driven by HELC and LE vacancies. This unmet skills demand tapered off in late 2022, which also coincided with the release of ChatGPT and the proliferation of generative AI tools. Thus, it is unclear how much of the subsequent decline in job vacancies was caused by technological advancement and how much by a post-pandemic labour market adjustment.

Job vacancies in HELC and LE occupations decreased by almost 50% from the fourth quarter of 2022 to the third quarter of 2025. Vacancies in HEHC occupations decreased by about 30% over the same period. The difference in the decrease in vacancies between HEHC and other occupations was statistically significant, but the difference in the decrease in vacancies between HELC and LE jobs was not These findings suggest that, so far, the unmet demand for skills potentially more exposed to and less complementary with AI has declined at a rate similar to the unmet demand for skills potentially less exposed to AI.

Data table for Chart 7

Data table for chart 7
Table summary
This table displays the results of Data table for chart 7 High exposure and high complementarity, High exposure and low complementarity and Low exposure, calculated using index units of measure (appearing as column headers).

	High exposure and high complementarity	High exposure and low complementarity	Low exposure
	index
Note .. not available for a specific reference period Notes: The dashed vertical line demarcates the beginning of mass availability of ChatGPT in 2022 Q4. Job Vacancy and Wage Survey data collection was suspended for the second and third quarters of 2020. Therefore, data for these reference periods are not available. The solid horizontal line demarcates the baseline of the job vacancy growth index. Any point above the line indicates a higher vacancy level relative to the one observed in 2022 Q4 and any point below the line indicates a lower vacancy level. All types of job vacancies are included. Job vacancy growth is expressed relative to the level observed in 2022 Q4. For example, an index of 110 would indicate 10% growth (110 minus 100) in job vacancies relative to the 2022 Q4 level, and an index of 90 would indicate that job vacancies were 10% lower (90 minus 100) relative to the 2022 Q4 level. Jobs were disaggregated based on the complementarity-adjusted artificial intelligence occupational exposure (C-AIOE) index developed by Felten et al. (2021) and Pizzinelli et al. (2023). An occupation has potentially high exposure if its AIOE index exceeds the median AIOE across all occupations; otherwise it is considered to have low exposure. Similarly, an occupation is considered to have potentially high complementarity if its complementarity parameter exceeds the median complementarity across all occupations; otherwise it is considered to have low complementarity. Sources: Statistics Canada, Job Vacancy and Wage Survey; and Occupational Information Network.
2015
Q1	41.0	49.4	53.2
Q2	39.7	51.4	45.9
Q3	39.5	46.5	43.0
Q4	39.0	45.5	40.0
2016
Q1	39.1	42.8	38.6
Q2	40.0	46.5	37.6
Q3	43.9	52.9	39.2
Q4	41.7	49.6	42.2
2017
Q1	40.4	51.4	45.8
Q2	45.1	52.2	48.5
Q3	46.9	55.4	51.2
Q4	49.8	58.5	54.8
2018
Q1	51.6	59.6	56.6
Q2	53.3	64.1	58.2
Q3	56.5	65.6	60.4
Q4	58.9	68.3	63.4
2019
Q1	59.5	66.9	62.4
Q2	59.8	65.1	61.3
Q3	58.8	63.7	61.5
Q4	59.2	63.0	59.0
2020
Q1	61.7	65.4	62.8
Q2	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period
Q3	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period
Q4	65.3	68.1	66.4
2021
Q1	67.6	71.7	69.0
Q2	75.6	81.3	77.9
Q3	85.5	95.8	107.9
Q4	92.5	104.1	111.8
2022
Q1	96.2	110.1	109.8
Q2	100.2	115.5	115.0
Q3	101.2	106.9	113.4
Q4	100.0	100.0	100.0
2023
Q1	97.3	89.6	94.1
Q2	95.2	84.7	87.6
Q3	91.3	76.3	81.5
Q4	88.1	72.2	74.2
2024
Q1	87.5	68.1	68.9
Q2	80.9	65.5	62.4
Q3	77.6	62.8	58.9
Q4	76.3	60.2	56.9
2025
Q1	74.4	60.8	55.4
Q2	72.4	55.7	54.1
Q3	70.8	53.5	53.2

Coding-intensive job vacancies increased over the 2021-to-2022 period at a faster rate than other vacancies (Chart 8). However, after the fourth quarter of 2022, coding-intensive vacancies declined more steeply than vacancies in other occupations. By the third quarter of 2025, vacancies for coding-intensive jobs requiring three years of experience or less fell by about 60%, while those for such jobs requiring more experience fell by about 30% (a statistically significant difference). Vacancies in other occupations requiring three years of experience or less decreased by roughly 45%, while those in such occupations requiring more experience decreased by about 20% (a statistically significant difference). However, it is difficult to say with any degree of certainty whether the steeper decline in coding-intensive vacancies reflects AI supplanting coding tasks or is instead the result of a natural correction following the rapid expansion in coding-intensive vacancies over the 2021-to-2022 period, which may have temporarily satisfied employers’ hiring needs. The drop in vacancies for coding-intensive jobs requiring less experience also aligns with the previous finding regarding the decrease in younger professionals in these occupations compared with their older counterparts. It is also possible that, as the pandemic-era boom in demand for coding-intensive roles subsided, the supply of young workers entering or retraining for these occupations declined, as the field became less attractive or perceived as offering fewer opportunities. However, given the relatively short timeline, this may be less applicable to degree holders—whose educational pathways and career decisions unfold over a longer period—but could be more relevant for those pursuing diplomas, certificates or shorter-term credentials, where training and entry into the labour market can occur over a shorter time.^Note 

Data table for Chart 8

Data table for chart 8
Table summary
This table displays the results of Data table for chart 8 Coding-intensive jobs requiring three years or less of experience, Coding-intensive jobs requiring more than three years of experience, Other jobs requiring three years or less of experience and Other jobs requiring more than three years of experience, calculated using index units of measure (appearing as column headers).

	Coding-intensive jobs requiring three years or less of experience	Coding-intensive jobs requiring more than three years of experience	Other jobs requiring three years or less of experience	Other jobs requiring more than three years of experience
	index
Note .. not available for a specific reference period Notes: The dashed vertical line demarcates the beginning of mass availability of ChatGPT in 2022 Q4. Job Vacancy and Wage Survey data collection was suspended for the second and third quarters of 2020. Therefore, data for these reference periods are not available. The solid horizontal line demarcates the baseline of the job vacancy growth index. Any point above the line indicates a higher vacancy level relative to the one observed in 2022 Q4 and any point below the line indicates a lower vacancy level. All types of job vacancies are included. The following 2021 National Occupational Classification occupations were selected as coding-intensive jobs: data scientists, cybersecurity specialists, business systems specialists, information systems specialists, database analysts and data administrators, computer systems developers and programmers, software engineers and designers, software developers and programmers, web designers, and web developers and programmers. Job vacancy growth is expressed relative to the level observed in 2022 Q4. For example, an index of 110 would indicate 10% growth (110 minus 100) in job vacancies relative to the 2022 Q4 level, and an index of 90 would indicate that job vacancies were 10% lower (90 minus 100) relative to the 2022 Q4 level. Jobs were disaggregated based on the complementarity-adjusted artificial intelligence occupational exposure (C-AIOE) index developed by Felten et al. (2021) and Pizzinelli et al. (2023). An occupation has potentially high exposure if its AIOE index exceeds the median AIOE across all occupations; otherwise, it is considered to have low exposure. Similarly, an occupation is considered to have potentially high complementarity if its complementarity parameter exceeds the median complementarity across all occupations; otherwise, it is considered to have low complementarity. Sources: Statistics Canada, Job Vacancy and Wage Survey; and Occupational Information Network.
2015
Q1	36.4	36.0	50.0	63.5
Q2	46.7	33.7	46.3	60.6
Q3	34.2	38.2	42.6	57.6
Q4	37.9	36.0	40.7	51.7
2016
Q1	38.6	30.5	38.6	50.8
Q2	41.0	34.2	40.8	51.1
Q3	48.2	42.1	42.9	51.8
Q4	42.5	37.5	44.8	53.3
2017
Q1	44.1	38.8	46.7	54.8
Q2	48.2	39.6	47.9	55.9
Q3	55.6	45.3	50.7	57.0
Q4	51.4	49.1	54.2	61.1
2018
Q1	55.1	53.0	55.5	61.5
Q2	61.9	56.8	57.9	63.4
Q3	69.6	52.3	59.8	70.3
Q4	67.4	52.1	62.2	73.4
2019
Q1	69.7	58.6	61.9	70.8
Q2	66.4	55.4	60.7	70.5
Q3	69.8	54.2	59.3	67.3
Q4	71.5	61.7	58.8	69.7
2020
Q1	77.7	69.4	60.3	69.5
Q2	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period
Q3	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period
Q4	68.8	72.5	67.1	67.1
2021
Q1	75.5	92.0	68.0	77.3
Q2	84.2	99.8	77.4	83.8
Q3	102.2	105.0	97.7	97.8
Q4	121.1	129.2	105.9	99.5
2022
Q1	121.0	129.2	106.4	104.8
Q2	151.2	125.9	111.1	107.5
Q3	113.7	114.1	106.7	104.1
Q4	100.0	100.0	100.0	100.0
2023
Q1	75.8	83.3	92.9	89.3
Q2	72.4	75.0	87.4	90.9
Q3	55.8	58.1	81.0	84.5
Q4	56.0	54.2	76.3	82.1
2024
Q1	56.7	59.7	69.8	83.1
Q2	40.7	60.8	65.6	81.5
Q3	42.9	63.1	61.2	83.1
Q4	41.7	63.9	58.6	83.7
2025
Q1	42.9	58.1	59.0	78.7
Q2	42.3	65.4	55.4	78.9
Q3	36.4	66.5	53.6	81.3

Jobs potentially more exposed to artificial intelligence are generally of higher quality than other jobs

Although job displacement is the central concern associated with technological disruption, new technologies can also affect the quality of jobs, such as by reducing wages by lowering the barriers to entry into certain professions. For example, global positioning systems have decreased the need for taxi or rideshare drivers to memorize maps. New technologies can also have a wage-increasing effect and lead to some occupations becoming more specialized. For example, the grammar and spell-check functionality of word processing software removed simple editing tasks for proofreaders, leading to these roles becoming more specialized, where they now help people improve their writing instead of simply correcting spelling and grammar.^Note 

In terms of some job quality features, HEHC jobs were more likely than HELC and LE jobs to be full-time, permanent or both (i.e., full-time and permanent); be high-tenured (10 years or more with the same employer); be unionized or covered by a collective agreement; be associated with a workplace pension plan; and have higher hourly wage rates. With the exception of unionization or collective agreement coverage and high tenure, HELC jobs were more likely to be associated with the aforementioned job quality features than LE jobs. This ranking remained the same over the last decade.

HEHC jobs averaged around $48.50 per hour in 2025, compared with $35.40 per hour for HELC jobs and $28.50 per hour for LE jobs (Table 1). Real hourly wages grew by about 5% from November 2022 to December 2025, with no significant difference in growth rates across the three occupation groups. In the longer term, over the last decade, HEHC hourly wage rates grew at a relatively faster pace. This is not surprising given that, on average, around 60% of HEHC employees had a bachelor’s degree or above over the last decade, compared with around 40% of HELC employees and about 15% of LE employees.

Around 91% of HEHC jobs were full-time in 2025, compared with 83% of HELC jobs and 78% of LE jobs. Around 91% of HEHC jobs were permanent, compared with 89% of HELC jobs and 86% of LE jobs. About 85% of HEHC jobs were both full-time and permanent, compared with 77% of HELC jobs and 70% of LE jobs.

Around 36% of HEHC jobs were unionized or covered by a collective agreement, compared with 33% of LE jobs and 23% of HELC jobs. The relatively higher unionization or collective agreement coverage rates among HEHC and LE employees are explained by the composition of jobs within these broad categories. HEHC jobs include a relatively large concentration of professional health care, education and government occupations, which traditionally have strong union representation. Jobs classified as LE include a substantial number of skilled trades jobs, many of which are also unionized. By contrast, the lower rate of unionization or collective agreement coverage among HELC employees suggests that workers in these roles may have fewer avenues for support or recourse in the event of AI-driven job losses.

For more than one-third (37%) of HEHC jobs, employees had been with the same employer for at least 10 years, compared with 25% of HELC jobs and 24% of LE jobs. About 66% of HEHC jobs were associated with workplace pension plans, compared with 52% of HELC jobs and 41% of LE jobs.

Coding-intensive professions averaged around $54.40 per hour in 2025, compared with $35.70 per hour in other occupations. Real hourly wages in both occupation groups grew by around 5% from November 2022 to December 2025, with no significant difference between their growth trajectories. Over a 10-year period, however, real hourly wages in coding-intensive professions grew at a faster pace (15%) than those in other jobs (9%). 

Nearly all coding-intensive jobs (98%) were full-time, compared with about 83% in other occupations. Coding-intensive jobs were also more likely to be permanent, compared with other jobs (95% versus 88%, respectively), and more likely to be both full-time and permanent (93% versus 76%, respectively). Coding-intensive jobs (60%) were also more likely to be associated with workplace pension plans than other jobs (51%). However, coding-intensive jobs were less likely to be high-tenured, compared with other jobs (23% versus 28%, respectively), and less likely to be unionized or covered by a collective agreement (14% versus 31%, respectively).

The change in occupational mix three years after the widespread availability of generative artificial intelligence is not markedly different from other periods of technological change

Chart 9 offers a historical perspective on how the change in occupational mix since the widespread availability of generative AI compares with other periods of technological change. Employing an adaptation of Duncan and Duncan’s (1955) dissimilarity index, as in Gimbel et al. (2025), Chart 9 illustrates the degree of compositional change in the distribution of occupations relative to major technology adoption or availability periods.^Note  The index is bounded by 0 and 100, with a higher value indicating greater dissimilarity in occupational composition relative to the reference period. Note that this index simply measures change over this period and does not show anything regarding the cause of that change. Changes can occur by workers switching jobs, workers losing jobs or unemployed people getting a new job—all of which can be caused by technological transformation or other economic factors.

In addition to calculating the dissimilarity index for AI with a reference month of November 2022, indices are also shown for the mass adoption of computers in the late 1980s, the spread of Internet technologies from the mid-1990s and the onset of the pandemic from April 2020, as well as a pre-pandemic baseline period of relative stability from 2016 to 2019. The reference months were chosen to align with the U.S. study by Gimbel et al. (2025). Chart 9 shows that 37 months following the release of ChatGPT (i.e., November 2022 to December 2025), roughly 8% of employees would have to change jobs to restore the original occupational distribution observed in November 2022. Although this is a non-negligible number, it is not markedly different than the numbers observed three years following the mass adoption of computers or of the Internet. Moreover, the change in occupational mix since November 2022 does not appear to be substantially different than that observed over the reference period from 2016 to 2019. This suggests that the shifts in occupational composition may have already been underway before the mass availability of generative AI. Unsurprisingly, the onset of the pandemic in April 2020 generated more immediate shifts in occupational composition compared with the advent of computers, the Internet and generative AI. This was mostly driven by the sudden decline of service jobs in restaurants, retailers and other similar establishments induced by lockdowns and other health measures during the pandemic.

Although not directly comparable with the U.S. study by Gimbel et al. (2025), the Canadian results generally align with it. The results suggest that despite concerns about a rapid transformation of the labour market in the wake of generative AI, the change in Canada’s occupational mix over the three years since the mass availability of generative AI is not markedly different from previous periods of technological disruption. The pace and magnitude of occupational reallocation following the advent of AI are comparable to those seen after the mass adoption of computers and the Internet.

It is important to note, however, that although generative AI is widely available to the general public, mass adoption of these technologies has not yet occurred in Canada. Recent data indicate that about 12% of Canadian businesses reported using AI in the delivery of goods or services (Bryan et al., 2025). One factor potentially limiting immediate widespread impact is the need for existing legacy systems and workflows to adapt to new AI technologies, which can slow integration and utilization. Consequently, the results depicted in Chart 9 may represent only the initial phase of technological diffusion. The full impact of generative AI on occupational dissimilarity remains to be seen. Therefore, the results from Chart 9 provide an alternative lens to assess whether generative AI has had a measurable effect on the occupational mix to date, while acknowledging that more widespread adoption—and further adaptation of legacy systems—in the future could potentially lead to greater changes in occupational composition.

Data table for Chart 9

Data table for chart 9
Table summary
The information is grouped by Months since reference date indicated in the legend (appearing as row headers), Artificial intelligence (reference November 2022), COVID-19 pandemic (reference April 2020), Baseline (reference June 2016), Internet (reference January 1996) and Computers (reference June 1987), calculated using percentage point change units of measure (appearing as column headers).

Months since reference date indicated in the legend	Artificial intelligence (reference November 2022)	COVID-19 pandemic (reference April 2020)	Baseline (reference June 2016)	Internet (reference January 1996)	Computers (reference June 1987)
	percentage point change
Note .. not available for a specific reference period Notes: Employment refers to the main job of employees aged 15 and older. The estimates presented above are based on an adaptation of Duncan and Duncan's (1955) dissimilarity index. It is calculated as the sum of the absolute differences in occupational shares in a given month and a reference month. This number is then multiplied by 50 (100 divided by 2) so that it can be interpreted as percentage point changes. The index indicates the percentage of employees in a given month who would have to change jobs to restore the occupational distribution observed in the reference month. A higher index indicates greater dissimilarity in occupational composition. This can occur by employees changing jobs, employees losing jobs or unemployed people getting new jobs. The index simply measures change in occupational mix and remains agnostic as to the reason for that change. The estimates presented here are based on six-month seasonally adjusted moving averages of employment across occupations classified at the five-digit level of the 2021 National Occupational Classification. This chart is based on a similar chart produced by Gimbel et al. (2025) for the United States. However, the estimates are not directly comparable because of the differences in data and occupational classification systems. The reference dates on this chart, except for the COVID-19 pandemic, refer to the approximate start date of the adoption and/or availability of the specified technology. Gimbel et al. (2025) used January 1984 as the reference date for mass adoption of computers. However, the Labour Force Survey did not have occupational information before January 1987. The reference date for the mass adoption of computers was therefore set to June 1987 (the first data point in the six-month moving average from January 1987 to June 1987). The baseline period was from June 2016 to December 2019, when the economy was relatively stable before the COVID-19 pandemic. The index for the COVID-19 pandemic covers the period from April 2020 to October 2022 only, since extending the index beyond October 2022 would conflate the potential effects of the mass availability of generative artificial intelligence, starting in November 2022. Source: Statistics Canada, Labour Force Survey, January 1987 to December 2025.
0	0.0	0.0	0.0	0.0	0.0
1	1.1	1.6	1.1	1.1	1.2
2	1.9	3.0	2.0	2.0	2.1
3	2.7	4.3	2.8	2.8	2.9
4	3.5	5.4	3.5	3.5	3.6
5	4.2	5.9	4.2	4.1	4.2
6	4.8	5.9	4.7	4.5	4.7
7	5.3	6.0	5.1	4.8	5.1
8	5.6	6.0	5.4	5.0	5.4
9	5.8	6.3	5.7	5.1	5.5
10	5.9	6.5	5.8	5.3	5.6
11	6.1	6.5	5.9	5.3	5.6
12	6.1	6.7	6.1	5.4	5.5
13	6.2	6.8	6.3	5.4	5.6
14	6.2	6.9	6.3	5.5	5.6
15	6.3	6.8	6.2	5.6	5.6
16	6.4	6.6	6.1	5.5	5.7
17	6.5	6.8	6.1	5.5	5.9
18	6.6	6.7	6.1	5.5	5.9
19	6.6	6.6	6.2	5.5	6.1
20	6.7	6.6	6.3	5.5	6.2
21	6.8	6.8	6.6	5.5	6.3
22	6.9	6.9	6.6	5.6	6.4
23	6.9	6.9	6.6	5.7	6.4
24	7.0	6.9	6.6	5.8	6.4
25	7.1	6.8	6.5	6.0	6.4
26	7.2	6.9	6.5	6.1	6.5
27	7.1	6.8	6.4	6.1	6.5
28	7.1	6.8	6.5	6.2	6.4
29	7.0	6.8	6.6	6.1	6.4
30	7.1	6.9	6.7	6.1	6.2
31	7.0	.. not available for a specific reference period	6.7	6.2	6.2
32	7.1	.. not available for a specific reference period	6.8	6.4	6.2
33	7.1	.. not available for a specific reference period	7.0	6.7	6.2
34	7.3	.. not available for a specific reference period	.. not available for a specific reference period	6.9	6.4
35	7.4	.. not available for a specific reference period	.. not available for a specific reference period	7.1	6.6
36	7.4	.. not available for a specific reference period	.. not available for a specific reference period	7.2	6.8
37	7.6	.. not available for a specific reference period	.. not available for a specific reference period	7.3	6.9
38	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.2	7.0
39	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.2	7.1
40	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.1	7.1
41	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.1	7.3
42	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.1	7.4
43	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	6.9	7.5
44	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	6.8	7.6
45	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	6.8	7.6
46	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	6.7	7.6
47	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	6.7	7.6
48	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	6.8	7.6
49	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	6.9	7.6
50	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.0	7.7
51	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.2	7.7
52	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.5	7.8
53	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.7	7.9
54	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	7.9	8.1
55	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	8.0	8.2
56	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	8.0	8.3
57	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	8.1	8.3
58	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	8.2	8.2
59	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	8.4	8.1
60	.. not available for a specific reference period	.. not available for a specific reference period	.. not available for a specific reference period	8.7	8.0

Conclusion

While AI is often associated with concerns about widespread job loss, early evidence from Canada since the mass availability of generative AI tools like ChatGPT suggests a more nuanced reality, highlighting a labour market evolving amid a confluence of technological, demographic and economic factors. Overall employment continued to grow three years after the mass availability of generative AI, with no clear evidence that jobs potentially more exposed to and less complementary with AI experienced disproportionate declines compared with other occupations. Although distributional effects within demographics, such as weaker job growth for younger and less educated workers, were observed, these may be due to multiple underlying economic factors.

Three years following the widespread availability of generative AI, the overall mix of occupations in the Canadian labour market does not appear to be markedly different from previous periods of technological disruption, such as the introduction of computers or the Internet. This suggests that the pace and magnitude of occupational reallocation in response to generative AI, while non-negligible, align with historical patterns of technological change rather than representing an unprecedented transformation.

Nonetheless, substantial uncertainty persists. The diffusion of AI is still in its early stages, with mass adoption yet to occur and the adaptation of legacy systems potentially delaying impacts. Moreover, multiple overlapping economic shocks—from the aftermath of the pandemic to rapid demographic shifts caused by increased immigration and to recent trade dynamics with the United States—make it challenging to disentangle the effects of AI on employment outcomes. Despite the lack of evidence for large-scale AI-induced job loss to date, the transformative potential of AI and uncertainty about longer-term consequences highlight the need for continued tracking and monitoring of the labour market.

References

Bryan, V., S. Sood and C. Johnston. 2024. Analysis on artificial intelligence use by businesses in Canada, second quarter of 2024. Analysis in Brief. Statistics Canada Catalogue no. 11-621-M. Ottawa: Statistics Canada.

Bryan, V., S. Sood and C. Johnston. 2025. Analysis on artificial intelligence use by businesses in Canada, second quarter of 2025. Analysis in Brief. Statistics Canada Catalogue no. 11-621-M. Ottawa: Statistics Canada.

Brynjolfsson, E., B. Chandar and R. Chen. 2025. Canaries in the coal mine? Six facts about the recent employment effects of artificial intelligence. Stanford University Working Paper.

Chandar, B. 2025. AI and labor markets: What we know and don’t know. Stanford University Digital Economy Lab Insights.

Dingel, J. I. and B. Neiman. 2020. How many jobs can be done at home? Journal of Public Economics 189.

Duncan, O. D. and B. Duncan. 1955. A methodological analysis of segregation indexes. American Sociological Review 20 (2): 210-217.

Eckhardt, S. and N. Goldschlag. 2025. AI and jobs: The final word (until the next one). Economic Innovation Group.

Eloundou, T., S. Manning, P. Mishkin and D. Rock. 2024. GPTs are GPTs: Labor market impact potential of LLMs. Science 384 (6702): 1306-1308.

Felten, E., M. Raj and R. Seamans. 2021. Occupational, industry, and geographic exposure to artificial intelligence: A novel dataset and its potential uses. Strategic Management Journal 42(12): 2195-2217.

Frenette, M. 2025. The changing nature of work in Canada: 1987 to 2024. Economic and Social Reports (February). Statistics Canada Catalogue no. 36-28-0001. Ottawa: Statistics Canada.

Frenette, M. and K. Frank. 2020. Automation and job transformation in Canada: Who’s at risk? Analytical Studies Branch Research Paper Series. Statistics Canada Catalogue no. 11F0019M. Ottawa: Statistics Canada.

Frey, C. B. and M. Osborne. 2013. The future of employment: How susceptible are jobs to computerisation? Oxford Martin Programme on Technology and Employment. Oxford: Oxford Martin School, University of Oxford.

Gimbel, M., M. Kinder, J. Kendall and M. Lee. 2025. Evaluating the impact of AI on the labor market: Current state of affairs. The Budget Lab. Yale University.

Government of Canada. 2025. Guide on the use of generative artificial intelligence [web page].

Gmyrek, P., J. Berg, K. Kamiński, F. Konopczyński, A. Ładna, B. Nafradi, K. Rosłaniec and M. Troszyński. 2025. Generative AI and jobs: A refined global index of occupational exposure. ILO Working Paper 140.

Hampole, M., D. Papanikolaou, L. D. W. Schmidt and B. Seegmiller. 2025. Artificial intelligence and the labor market. NBER Working Paper No. 33509.

Humlum, A. and E. Vestergaard. 2025. Large language models, small labor market effects. NBER Working Paper No. 33777.

Kochhar, R. 2024. Which U.S. workers are more exposed to AI on their jobs? Pew Research Center.

Li, V. and G. Dobbs. 2025. Right brain, left brain, AI brain: AI’s impact on jobs and skill demand in Canada’s workforce. The Dais.

Mehdi, T. and M. Frenette. 2024. Exposure to artificial intelligence in Canadian jobs: Experimental estimates. Economic and Social Reports (September). Statistics Canada Catalogue no. 36‑28-0001. Ottawa: Statistics Canada.

Mehdi, T. and R. Morissette. 2021. Working from home: Productivity and preferences. StatCan COVID‑19: Data to Insights for a Better Canada. Statistics Canada Catalogue no. 45-28-0001. Ottawa: Statistics Canada.

Mehdi, T. and R. Morissette. 2024. Experimental estimates of potential artificial intelligence occupational exposure in Canada. Analytical Studies Branch Research Paper Series. Statistics Canada Catalogue no. 11F0019M. Ottawa: Statistics Canada.

Nedelkoska, L. and G. Quintini. 2018. Automation, skills use and training. OECD Social, Employment and Migration Working Papers, No. 202, Paris: OECD Publishing.

Oschinski, M. and R. Walia. 2025. Harnessing generative AI: Navigating the transformative impact on Canada’s labour market. Institute for Research on Public Policy, Study No. 97.

Oschinski, M. and R. Wyonch. 2017. Future shock? The impact of automation on Canada’s labour market. C.D. Howe Institute, Commentary No. 472.

Pizzinelli, C., A. J. Panton, M. M. Tavares, M. Cazzaniga and L. Li. 2023. Labor market exposure to AI: Cross-country differences and distributional implications. IMF Working Paper no. 216. Washington, D.C.: International Monetary Fund.

Statistics Canada. 2024a. Table 33-10-0825-01 Use of artificial intelligence by businesses and organizations in producing goods or delivering services over the last 12 months, second quarter of 2024. Ottawa: Statistics Canada.

Statistics Canada. 2024b. Table 33-10-0826-01 Extent to which artificial intelligence has reduced tasks previously performed by employees and the impact of artificial intelligence use on total employment, second quarter of 2024. Ottawa: Statistics Canada.

Statistics Canada. 2025a. Table 33-10-1004-01 Use of artificial intelligence by businesses and organizations in producing goods or delivering services over the last 12 months, second quarter of 2025. Ottawa: Statistics Canada.

Statistics Canada. 2025b. Table 33-10-1006-01 Extent to which artificial intelligence has reduced tasks previously performed by employees and the impact of artificial intelligence use on total employment, second quarter of 2025. Ottawa: Statistics Canada.

Statistics Canada. 2025c. Job vacancies, second quarter 2025. The Daily. Ottawa: Statistics Canada.

Statistics Canada. 2025d. Table 17-10-0-0005-01 Population estimates on July 1, by age and gender. Ottawa: Statistics Canada.