Legislators in the European Union granted approval to the 27-nation bloc's artificial intelligence law on Wednesday, paving the way for what it describes as its world-leading regulations to come into force later this year.
The AI Act, which aims to ensure safety, fundamental rights, and innovation, received significant support from lawmakers in the European Parliament, with 523 votes in favor, 46 against, and 49 abstentions.
The new law is said to serve as a model for other governments worldwide navigating the complexities of regulating the rapidly advancing technology, reported The Associated Press.
"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 Romanian lawmaker who was a co-leader of the Parliament negotiations on the law. "The AI act is not the end of the journey but the starting point for new governance built around technology."
First proposed five years ago, the legislation sets differing levels of risk and creates a new oversight office to regulate the development and use of artificial intelligence models.
Low-risk systems, like content recommendation or spam filters, face minimal regulations, such as the need to disclose they are powered by AI, while high-risk AI applications, such as those used in critical infrastructure, education, healthcare, law enforcement, border management, and elections, will be subject to stringent requirements.
Certain uses of AI will be prohibited, due to their deemed unacceptable risks, such as for social scoring systems and certain surveillance methods like facial recognition for public use by law enforcement, except in cases of serious crimes.
Early drafts of the law had focused on limited AI tasks, such as for resume scanning for job applications, but the emergence of general-purpose AI models, like OpenAI's ChatGPT, prompted EU policymakers to adapt by including provisions for generative AI models, requiring developers to disclose data sources and comply with copyright law, especially for deepfake content.
Enza Iannopollo, principal analyst at leading research and advisory company Forrester, was quoted by the BBC as saying the law would make the EU the "de facto" global standard for trustworthy AI.
"The adoption of the AI Act marks the beginning of a new AI era and its importance cannot be overstated," she said. "The EU AI Act is the world's first and only set of binding requirements to mitigate AI risks."
Brussels will establish an AI Office that will enforce the law for general-purpose AI systems, with potential fines of up to 35 million euros ($38 million) or 7 percent of a company's global revenue for any AI Act violations.
