The European Union is pushing ahead with plans to regulate artificial intelligence.
The European Parliament passed the artificial intelligence law on Wednesday with 523 votes in favor, 46 against and 49 abstentions.
“Europe is now the global standard for artificial intelligence,” EU Internal Market Commissioner Thierry Breton told X. “We regulate as much as we can, but only as much as necessary.”
It's the first time a major regulator has fully expanded AI controls to protect people from the technology's potential risks. Other countries, including China, have already passed laws on specific uses of artificial intelligence.
The bill has been questioned by some commentators, including artificial intelligence expert and deepfake Henry Ajdel, who said the bill is “too ambitious.” He called the move a positive move overall but warned that it would weaken Europe's global competitiveness.
"My concern is that companies will explicitly avoid development in certain areas where there are strict and comprehensive regulations," he told Business Insider. We will see some countries act as tax havens for AI policy, explicitly avoiding the implementation of strict laws aimed at attracting certain types of organizations.
This measure has been implemented for some time. It was first discussed in 2021 and provisionally agreed in negotiations with member states in December 2023.
EU legislation divides the risks of AI apps into three categories and apps that pose unacceptable risks are banned.
Special legal requirements apply to high-risk applications classified in the second category, while applications in the third category are largely unregulated.
KEY MILESTONE
Neil Silver, CEO of California-based CalypsoAI, said: "This legislation contains complex and potentially costly compliance requirements that place an initial burden on companies and provide opportunities for more responsible and transparent AI development."
He called the law "an important milestone in the development of artificial intelligence" and offers companies the opportunity to sooner or later incorporate social values into their products.
Depending on final testing, the regulation is expected to go into effect in May. The new rules will be introduced gradually from 2025.
Exactly how these rules apply to companies is still relatively unclear.
Avani Desai, CEO of cybersecurity company Shellman, said the law could have similar effects to the EU's General Data Protection Regulation (GDPR), which requires US companies to meet certain requirements to operate in Europe.
As the EU Commission creates an artificial intelligence office and begins setting standards, companies unsure of the rules can expect more details on specific requirements in the coming months, according to Marcus Evans of law firm Norton Rose Fulbright.
"The first obligations of the AI Act come into effect this year, and the rest over the next three years, so companies need to prepare as quickly as possible to avoid falling foul of the new rules," he said.
0 Comments