The European Parliament approved provisional AI legislation intended to prevent exploitation of the technology and protect fundamental rights, as the continent edges closer to adopting rules as law.

The green light for the AI Act comes after a provisional agreement was struck between the European Council, European Parliament and the European Commission in December 2023, following three days of negotiations.

In an announcement, the Parliament stated the AI regulation was endorsed by the majority of its members with 523 votes in favour, 46 against and 49 abstentions.

Brando Benifei, co-rapporteur of the internal market committee at the European Parliament, described Europe’s AI legislation as “the world’s first binding law” on the technology, pointing to objectives of the legislation around combatting discrimination and encouraging transparency and opportunities around usage.

Benifei further credited the Parliament for the impending ban of “unacceptable AI practices”, a move expected to protect worker and citizen rights.

The law is now subject to a “final lawyer-linguist check” and the Parliament noted it still needs to be “formally endorsed” by the EC. Countries in the bloc are expected to give formal approval in May.

With the legislation set to enter into force early next year, European Commissioner for internal markets Thierry Breton claimed Europe “is now a global standard-setter in AI”.

The AI Act includes provisions that will lead to the banning of biometrics “based on sensitive characteristics”, emotional and behavioural recognition and predictive policing. However, the use of biometrics to investigate “criminal offences” will be exempt.