The European Union on Wednesday took a major step toward passing what would be one of the first major laws to regulate artificial intelligence, a potential model for policymakers around the world struggling to put guardrails on the future. rapidly developing technology.
The European Parliament, a key legislature of the EU, has passed a draft bill known as the AI Bill, which would place new restrictions on what are perceived to be the technology’s riskiest applications. It would severely limit the use of facial recognition software, while makers of AI systems like the ChatGPT chatbot would have to reveal more about the data used to create their programs.
The vote is one step in a longer process. A final version of the law is not expected to be passed until later this year.
The European Union is further ahead than the United States and other major Western governments in regulating AI. The 27-nation bloc has been debating the topic for more than two years, and the issue took on a new urgency following the release of ChatGPT last year, which raised concerns about the potential effects of the technology on employment and society.
Policymakers everywhere from Washington to Beijing are now racing to take control of an evolving technology that alarms even some of the earliest makers. In the United States, the White House has released policy ideas that include rules for testing AI systems before they are publicly available and for protecting privacy rights. In China, design rules unveiled in April would require chatbot makers to abide by the country’s strict censorship rules. Beijing is also taking more control over how makers of AI systems use data.
How effective any regulation of AI can be is unclear. In a sign that the technology’s new capabilities are seemingly emerging faster than lawmakers can address, previous versions of EU law didn’t pay much attention to so-called generative AI systems like ChatGPT, which can produce text, images and video in response to directions.
In the latest version of the European bill passed on Wednesday, generative AI would face new transparency requirements. That includes publishing summaries of copyrighted material used to train the system, a proposal supported by the publisher but opposed by tech developers because it’s technically infeasible. Creators of generative AI systems should also take precautions to avoid generating illegal content.
Francine Bennett, acting director of the Ada Lovelace Institute, a London-based organization that has pushed for new AI laws, said the EU proposal was an “important milestone”.
“Fast-moving and rapidly reusable technology is of course difficult to regulate, if even the companies building the technology are not completely clear on how things will turn out,” Ms Bennett said. “But it would certainly be worse for all of us to continue to operate without some adequate regulation.”
The EU bill takes a “risk-based” approach to regulating AI, focusing on applications with the greatest potential for human harm. This includes where AI systems are used to operate critical infrastructure such as water or energy, in the legal system and in determining access to public services and government benefits. Creators of the technology will have to conduct risk assessments before using the technology on a daily basis, similar to the drug approval process.
A tech industry group, the Computer & Communications Industry Association, said the European Union should avoid overly broad regulations that hinder innovation.
“The EU will become a leader in regulating artificial intelligence, but it remains to be seen whether it will lead in AI innovation,” said Boniface de Champris, the group’s Europe policy manager. “Europe’s new AI rules should effectively address well-defined risks, while leaving enough flexibility for developers to deliver useful AI applications that benefit all Europeans.”
An important point of discussion is the use of facial recognition. The European Parliament voted to ban the use of live facial recognition, but questions remain over whether exemptions should be allowed for national security and other law enforcement purposes.
Another provision would prohibit companies from scraping biometric data from social media to build out databases, a practice that came under scrutiny after being used by facial recognition company Clearview AI.
Tech leaders have tried to influence the debate. Sam Altman, the CEO of OpenAI, the creator of ChatGPT, has visited at least 100 US legislators and other global policymakers in South America, Europe, Africa and Asia in recent months, including Ursula von der Leyen, President of the European Commission. Mr Altman has called for AI to be regulated, but has also said the EU’s proposal may be prohibitively difficult to implement.
After Wednesday’s vote, a final version of the law will be negotiated between representatives of the three branches of the European Union: the European Parliament, the European Commission and the Council of the European Union. Officials said they hope to reach a final agreement by the end of the year.
