The system demands machines make moral calls in real time.
A major expansion of the UK’s Online Safety Act (OSA) has taken effect, legally obliging digital platforms to deploy surveillance-style systems that scan, detect, and block user content before it can be seen.
The government’s new Online Safety Act 2023 (Priority Offenses) (Amendment) Regulations 2025, which came into force on January 8, 2026, designates “cyberflashing” and “encouraging or assisting serious self-harm” as priority offenses, categories that trigger the strictest compliance duties under the OSA.
This marks a decisive move toward preemptive censorship. Services that allow user interaction, including messaging apps, forums, and search engines, must now monitor communications at scale to ensure that prohibited content is automatically filtered or suppressed before users can even encounter it.
To meet the law’s demands, companies are expected to rely heavily on automated scanning systems, content detection algorithms, and artificial intelligence models trained to evaluate the legality of text, images, and videos in real time.
The UK Department for Science, Innovation and Technology (DSIT) unveiled the changes through a promotional video showing a smartphone scanning AirDropped photos and warning the user that an “unwanted nude” had been detected.