There’s a Growing Annoyance in the American Workplace. It Could Be Coming for Your Job Next.

Erin Nash was a hospital chaplain whose job was to be with people in some of their worst moments, praying, holding hands, even singing with them. Shadowing her on her rounds, I watched as she managed to create brief peaceful moments with suffering patients and their families again and again, making temporary sanctuary between the thin blue medical curtains despite the buzzing alarms, fluorescent lighting, and constant stream of footsteps on the linoleum floors nearby.

I was surprised to learn that in addition to consoling the bereaved and calming the anxious, Erin (the names in this piece have been changed) had to fill out three separate charts—including the standard electronic health records system that many clinicians use—for every person she visited. She even carried around a cheat sheet to help her remember the codes, murmuring, under her breath, “Asking for a prayer is a resource, family together is a resource,” while she hunted and pecked at the keyboard. Nobody was being billed for Erin’s work, so why was she charting in triplicate?

To find the answer, I spent five years talking to workers like Erin, as well as the managers and engineers who are trying to design and impose the systems that control her work. Ultimately, the spread of data analytics into feeling labor is more than just the latest frontier in an inexorable drive toward increasing efficiency everywhere. It has implications for A.I., the future of work, and the stratification of human contact.

Slate receives a commission when you purchase items using the links on this page. Thank you for your support.

Erin is one of millions, from teachers to therapists to managers to hairdressers, whose work relies on relationship. By some accounts, the U.S. is moving from a “thinking economy” to a “feeling economy,” as many deploy their emotional antennae to bear witness and reflect back what they understand so that clients, patients, and students feel seen. I’ve come to call this work “connective labor,” and the connections it forges matter. It can be profoundly meaningful for the people involved, and it has demonstrable effects: We know that doctor–patient relationships, for instance, are more effective than a daily aspirin to ward off heart attacks.

But this work is increasingly being subjected to new systems that try to render it more efficient, measurable, and reproducible. At best, firms implement these systems assuming that such interventions will not get in the way of workers and clients connecting. At worst, they ignore or dismiss those connections altogether. Even these complex interpersonal jobs are facing efforts to gather information and assessment data and to introduce technology. Moneyball has come for connective labor.

Teachers have long complained about standardized testing and surveillance, but it is not just teachers anymore. I listened to scores of workers across many fields talk about the spread of counting and scripting, like Valerie Clausen, a hairdresser in Virginia. “They set a timer and they tell you, ‘OK, you can only talk for this long, and if your service time is slowing down, then stop the talking and get to doing what you need to do.’ You know, sometimes you need to stop and make real live contact with somebody, not just through the mirror, if you’re into that deep of a conversation. And they really don’t want that to happen. Everything put together is 22 minutes. That’s all you got. It doesn’t matter what. Twenty-two minutes.”

We can certainly have sympathy for the goals underlying these changes. Research finds that checklists and manuals can confer greater legitimacy upon many kinds of service work and hedge our bets against incompetence and discrimination, while also protecting workers from demanding or chaotic situations and clients. In part, the new systems aim for a society in which getting a good teacher or doctor or hairdresser is less dependent on being lucky or affluent.

Yet research also shows that the scripting of interactive service work threatens creativity and autonomy, transforms clients or patients into standardized “industrial objects,” and demoralizes workers, alienating them from their own feeling. More than 50 percent of physicians say they are burned out by data entry. In a Gallup survey, about half of teachers and nurses reported experiencing high levels of job-related stress. Changes in the way we organize connective labor are extracting enormous costs.

If there is a battle over the systematization of feeling work, we may have already conceded it when it comes to teachers and low-wage service workers like Valerie. But scripting and counting is invading other relational professional work, even therapy, perhaps the last redoubt of sentiment. When Sarah Merced, a therapist at a Veterans Affairs hospital, talked about the mandated treatment plans decreed from above, I was struck by her wistfulness. “I guess my wish as a clinician would be to be able to use some actual clinical thinking,” she said. “I went to school for many years to kind of figure out what best treatments work for most people, and so it would be lovely to be able to choose when to use these tools and when not, versus having them mandated.”

The industrial model has become so dominant, even in these spaces where deep feeling has profound impact, that it is something of a cultural juggernaut, shaping how people think and talk about value. In my research, I watched as even connective labor practitioners found themselves trapped within its interpretive framework, forced to articulate their contributions and constraints, however reluctantly, with its alien grammar.

This sort of influence was evident when I talked to Alan Krupner, Erin’s supervisor and the head of the chaplain group. The most important reason for all the charting that Erin did, Alan said, was that someday the hospital was going to decide the future of the chaplain service, and data supporting what they did would help chaplains steer their own future. “I felt like chaplains needed to learn how to count something because we live in a world increasingly driven by evidence. The question is, are we going to be the ones who say these things, or is someone else going to tell us that ‘we drive the bus, so you’re going to do what we tell you to’?” The chaplains might eventually be able to convince health care administrators that their work is important. “But to do that,” he said, “we need to get over ourselves.”

Alan’s argument is one thing—a statement about the inevitability of quantitative data deciding the place of a spiritual team in the hospital, attesting to medicine’s institutional muscle, as well as the intellectual hegemony of evidence-based thinking in his world. But his language—chaplains need “to learn how to count something” and “we need to get over ourselves,” with its casual mockery of chaplains as number-averse feelers—is another: evidence of a certain cultural supremacy of standardization.

Finally, the triumph of data in connective labor is more than the introduction of efficiency into the messy, chaotic world of human relations, akin to how statistics revolutionized talent-spotting in baseball. The degradation of these jobs—their reduction from artisanal emotional craftwork to a set of easily reproducible steps—actually acts as a slippery slope to their automation through A.I. and apps like Woebot, a therapy app, or Traitify, the personality test that McDonald’s uses in hiring. Moreover, because of the overlap of these trends with social inequality, we are approaching a world in which the affluent pay for personal connective labor from less-advantaged others, who might themselves have to get it delivered by app or A.I.

I am not advocating eliminating all standards or systems, procedures that some of my interviewees found useful. But we should pay close attention to the impact systems have on connective labor, the engine powering the outcomes we value, like learning and healing. One solution might be to protect feeling jobs from the data collection that threatens to overwhelm them, such as that plaguing primary care physicians. One medical assistant I spoke to acted as a scribe and treasured having more responsibility, but the real advantage was in the relationships she witnessed, she said. “The provider takes all their time looking at the patient, you know, ‘I’m here, I’m talking to you, I’m listening to what you’re saying.’ ”

Connective labor is increasingly being subjected to new systems that try to make it more predictable, measurable, efficient—and reproducible. If we continue to prioritize efficiency over relationship, we degrade jobs that have the potential to forge profound meaning between people and, along the way, make them more susceptible to automation and A.I., creating a new kind of haves and have-nots: those divided by access to other people’s attention.