The European Union has introduced the world’s most detailed AI regulations, extending protections to include health and psychological impacts while tightening transparency and oversight requirements for high-risk systems.
Europe is moving ahead with some of the world’s most detailed rules on artificial intelligence as policymakers focus not only on fraud and misinformation but also on the possibility that AI could damage users’ health. The European Union’s AI Act, described by the European Commission as the first comprehensive legal framework for AI, has been introduced in stages since 2024 and a key set of restrictions took effect on August 2 for high-risk systems used in regulated products such as medical devices, machinery and toys.
At the centre of the law is a risk-based model that divides AI into four categories: unacceptable risk, high risk, limited risk and minimal risk. According to the European Commission, the point is to ensure that AI systems contribute to society without creating outcomes that undermine safety or fundamental rights. The new rules also tighten requirements for transparency, human oversight and risk management in areas such as healthcare, where AI tools must be designed to support safe use rather than replace professional judgement.
Among the practices now prohibited are systems that use manipulation, deception or exploitation to steer behaviour in ways that could cause significant harm. The Commission’s recital on the ban specifically refers to subliminal techniques and other methods that can impair autonomy and decision-making. Industry guidance on the new framework says the law is intended to reduce risks such as fraud, misinformation and exploitation, while requiring companies to label AI-generated content and face enforcement measures if they fail to comply.
Doctors and mental health specialists say the concern is not limited to obvious abuse. Some warn that repeated exposure to highly persuasive or overstimulating AI systems could affect stress responses, sleep and longer-term decision-making. Nathan Carroll, a psychiatrist and national medical director at InSite Health, told The Post that subtle manipulation is especially troubling because the harm can be difficult to see. Alexander Acosta, a physician with experience in primary, emergency and intensive care medicine in Ecuador’s public health system, said the physical effects can resemble the stress reactions seen in acute medical emergencies.
The medical framework for identifying AI-related harm is still underdeveloped, however. Acosta said clinicians can recognise symptoms such as high blood pressure and sleep problems, but there is not yet a standard way to connect them directly to AI exposure. Carroll argued that regulation alone will not be enough and said public education and stronger protections for children are also needed. The European Union’s approach may therefore be a test case for whether AI regulation can address harms that are psychological, behavioural and potentially physical at the same time.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
