Europe turns to existing laws to regulate cutting-edge AI

• The European Union has been among the first to react to the concerns regarding AI, though there are no current laws on the books to help govern the use of the technology.
• The EU has long led the U.S. in regulating the tech industry and "there's no doubt" the bloc is outpacing the U.S. in addressing AI, Justin Sherman, a nonresident fellow at the Atlantic Council's Cyber Statecraft Initiative, told Axios.

State of play: EU regulators approved the AI Act last December, which is set to lay out comprehensive guidance governing AI tools, facial recognition and biometric surveillance.

• The law will be one of the first major regulatory frameworks for AI, Sherman said.
• The European Parliament is slated to vote on the legislation in June, after which final talks between the parliament and EU member states will begin.
• The aim is to close the deal by the end of 2023, but the law won't be implemented for roughly another two years, a grace period allowing companies and organizations to comply with the law, Boniface de Champris, Policy Manager at the Computer and Communications Industry Association’s Brussels office told Axios.

Meanwhile, national data protection authorities are turning to laws already on the books and applying them to AI.

• Italy's data protection agency temporarily banned ChatGPT in March and initiated a probe into a suspected breach of privacy concerns. The agency used provisions of the EU's privacy law — the General Data Protection Regulation (GDPR) — in order to crack down on ChatGPT, Reuters reported.
• Spain's data protection watchdog announced in April it would also launch an investigation into data breaches by ChatGPT and requested that the EU's privacy watchdog discuss concerns regarding the service.
• The same month, France's data protection watchdog CNIL announced that it was also investigating complaints about ChatGPT.
• The EU's privacy watchdog announced in April the creation of a task force on ChatGPT that would "foster cooperation and to exchange information on possible enforcement actions conducted by data protection authorities."

The big picture: Relying on existing laws is imperfect. The EU's forthcoming AI Act — once it is finalized — will hopefully address some of the legal gaps that remain unaddressed, Iverna McGowan, the director of the Center for Democracy and Technology's Europe office, told Axios.

• For instance, the EU has robust laws guaranteeing equal access to employment. But with the increased use of AI in recruitment processes, it's not currently clear in practice who is accountable for potential discrimination in the process — the producers of AI or the employers themselves, McGowan explained.
• "Under EU law, the burden of proof is actually on the employer to disprove discrimination, but how can they do that if they themselves don't understand how the decision was made?" she asked.

• If the AI Act includes sufficiently robust transparency provisions, people will be able to take discrimination concerns to court with the data or information to back up their claim, she added.

Zoom out: While the U.S. has nearly no AI-specific regulations on the books, the Biden administration unveiled a series of initiatives on Tuesday around the research, development and deployment of "responsible AI."

• At a U.S. Senate Judiciary subcommittee hearing last week, OpenAI CEO Sam Altman urged Congress to enact regulations for AI, floating the idea of an international body to help set standards and monitor AI.
• Multiple lawmakers at the hearing highlighted the Congress' previous failure to take early action to set rules for social media — a mistake they're determined not to repeat, Axios' Ashley Gold writes.
• A bipartisan group of senators has begun discussions about crafting a comprehensive AI bill.
• Yet none of the actions taken by the U.S. have "as much teeth" as the AI Act's regulation will have in Europe, McGowan said.

Our thought bubble, from Axios' Ashley Gold: Regulators around the world are used to the exercise of trying to apply old rules to new tech, sometimes successfully and sometimes not.

• But AI's fast pace provides an even bigger challenge and new, thorny policy issues.