The European Parliament (EP) plenary on Wednesday, the 13th of March adopted a landmark bill on artificial intelligence (AI) by 523 votes to 46 with 49 abstentions, making Europe the first major political organisation to legislate on the fast-developing industry, reports Politico.
The set of rules will ban some applications of AI, such as algorithmic social scoring and unethical practices, and tighten the rules for other AI systems like ChatGPT, that are
used in “high-risk” situations such as education, work or critical infrastructure management.
The AI rulebook requires “general purpose” AI models to comply with mitigation and testing obligations, which will be monitored and enforced by a newly created ” AI Office” within the European Commission.
The draft law was first drafted in 2021, before the rise of AI tools such as OpenAI’s ChatGPT, and was hampered by tensions between European legislators and European Union (EU) governments, including France and Germany, over concerns that strict regulation could hinder Europe’s progress in the global race for AI leadership.
Other points of disagreement included the possibility of a complete ban on facial recognition technology, which is limited but not banned in the final text, and the structure of the “AI Office” itself.
Thierry Breton, head of the European Commission’s internal market, said after the vote that
“Europe has become the global standard setter in trustworthy AI”.
The EU Council is expected to formally adopt the text on the use of AI in April, with the ban aspects coming into force at the end of 2024, and the rules on general purpose AI coming into force in the early 2025. The remaining regulations are expected to apply in 2026.
In the meantime, the Breton-sponsored “AI Pact” will help companies prepare for compliance.
Also read: EU Parliament to sue the Commission over Hungary funds
Follow us on Facebook and X!