E.U. Takes Major Step Toward Regulating A.I.

The European Union on Wednesday took an important step towards passing legislation that will be one of the first major laws to regulate artificial intelligence. This provides a potential model for policy makers around the world grappling with how to put guardrails on rapidly evolving technologies.

The EU’s main legislative body, the European Parliament, has passed a bill known as the AI ​​Law that places new limits on what is considered the riskiest uses of technology. It would slash the use of facial recognition software while requiring makers of AI systems like ChatGPT chatbots to disclose more about the data used to create their programs.

Voting is one step in a longer process. The final version of the law is not expected to be enacted until later this year.

The European Union is more advanced than the US and other Western countries when it comes to regulating AI.The bloc of 27 countries has been debating the subject for more than two years, and the release of ChatGPT last year, which heightened concerns, has led to this trend. The question is about the potential employment and social impact of this technology, which has taken on new urgency.

Policymakers from Washington to Beijing now race to control an evolving technology that worries even some of its early developers.In the United States, the White House announced policy ideas It includes rules for testing AI systems and protecting privacy rights before releasing them to the public. In China, draft rules released in April require chatbot makers to follow the country’s strict censorship rules. The Chinese government is also tightening control over how AI system makers use data.

It’s unclear how effective AI regulation will be. In a sign that new features of the technology are emerging faster than lawmakers can react, previous versions of EU law paid little attention to so-called generative AI systems like Chat GPT. ChatGPT can generate text, images and videos on demand. Display a prompt.

Generative AI will face new transparency requirements in the latest version of the European bill passed on Wednesday. This involves publishing summaries of copyrighted material used to train the system, a proposal supported by the publishing industry but technically Opposed as impracticable. Manufacturers of generative AI systems also need to take safeguards to prevent the generation of illegal content.

Francine Bennett, deputy director of the Ada Lovelace Institute, a London group pushing the new AI law, called the EU’s proposal a “major milestone”.

“It’s naturally difficult to regulate rapidly evolving, rapidly reusable technology. Even the companies building the technology aren’t entirely clear how things will unfold,” Bennett said. said. “But it would definitely be bad for all of us to continue operating without proper regulation at all.”

The EU bill takes a “risk-based” approach to AI regulation, focusing on applications that are most likely to harm humans. This includes places where AI systems are used to operate critical infrastructure such as water and energy, within legal systems, and in determining access to public services and government benefits. Manufacturers of this technology should conduct risk assessments before using the technology routinely, similar to the approval process for pharmaceuticals.

The Computer and Telecommunications Industry Association, a high-tech trade group, said the European Union should avoid overly broad regulations that stifle innovation.

“The EU is poised to become the leader in artificial intelligence regulation, but it remains to be seen whether it will take the lead in AI innovation,” said Bonifas de Champry, European policy manager at the group. “Europe’s new AI rules should effectively address well-defined risks while leaving enough flexibility for developers to deliver useful AI applications that benefit Europe as a whole.”

One of the main areas of discussion is the use of facial recognition. The European Parliament has passed a ban on the use of live facial recognition, but questions remain about whether exemptions should be allowed for national security or other law enforcement purposes.

Another provision bans companies from collecting biometric data from social media to build databases, a practice used by facial recognition company Clearview AI that came under heavy scrutiny.

Tech industry leaders are trying to influence the debate. Sam Altman, chief executive of OpenAI, developer of ChatGPT, has spent the last few months working with at least 100 US lawmakers, including European Commission President Ursula von der Leyen, as well as South American, European, He met with other global policy makers from Africa and Asia. . Altman calls for regulation of AI, but said it may be prohibitively difficult to comply with the EU’s proposals.

After Wednesday’s vote, the 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 European Union Council. Officials said they hope to reach a final deal by the end of the year.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button