Any product that uses artificial intelligence needs to be safety assessed for its entire lifespan under new rules that came into effect across the European Union on Dec. 13. Experts also warn that companies that use AI and tailor it to suit their own needs could technically be classed as “manufacturers” under the legislation and face the same duty of care as the tech firms that developed it.

The EU General Product Safety Regulation (GPSR) represents the most comprehensive overhaul and update of the bloc’s product safety framework in more than 20 years, experts say, as the legislation covers how products may evolve over time.

“The GPSR is clear that safety needs to be assessed across a product’s entire lifespan, including how a product is updated or changes over that lifespan,” warned Aonghus Heatley, director in the regulatory group at law firm Fieldfisher.

From now on, product safety assessment will need to cover areas including mental health; cybersecurity; the impact of software updates/software subsequently downloaded on products; AI and machine learning functionalities; and the different impacts these products might have on health and safety for different genders, as well as the risks they pose for vulnerable consumers.