The European Union has taken a big step in the regulation of artificial intelligence (AI) by granting final approval to the Artificial Intelligence Act. This pioneering legislation, which received overwhelming support from lawmakers in the European Parliament, is poised to take effect later this year, establishing a world-leading framework for other governments to follow in regulating rapidly evolving AI technologies.
"The AI Act has nudged the future of AI in a human-centric direction, in a direction where humans are in control of the technology and where it — the technology — helps us leverage new discoveries, economic growth, societal progress and unlock human potential," said Dragos Tudorache, a key figure in the Parliament's negotiations on the draft law.
The AI Act introduces a risk-based approach to regulation, with stringent scrutiny for high-risk AI applications such as medical devices and critical infrastructure. It also prohibits certain AI uses deemed to pose unacceptable risks, including social scoring systems and emotion recognition in schools and workplaces.
In response to the rapid advancement of generative AI models like OpenAI's ChatGPT, the Act has been updated to include provisions for these technologies. Developers of generative AI models will now need to provide detailed summaries of the data used to train their systems and comply with EU copyright law.
Industry leaders have expressed their views on the new regulations. "The EU AI law was put in place to help ensure AI technologies are developed and used in a way that is safe, respects EU laws and values, and upholds fundamental rights," said Jadee Hanson, CISO of Vanta and former CISO of Target. "There is a clear benefit to integrating AI technology into the many products we use today. The EU AI Act functions as a necessary guardrail, curtailing AI applications that should be off-limits."
The AI Act is expected to officially become law by mid-2023, with provisions starting to take effect in stages. Each EU country will establish its own AI watchdog to enforce the regulations, and violations could result in fines of up to 35 million euros or 7% of a company's global revenue.
As the world's first comprehensive set of AI rules, the EU's AI Act marks a significant milestone in the governance of emerging technologies and sets the stage for further AI-related legislation in the future.