US presses G20 for AI deregulation as EU vows to enforce AI Act

Al Jazeera reports that the United States used a Group of 20 ministerial meeting on innovation, held on a Tuesday in Chapel Hill, North Carolina, to argue against AI-specific regulation, while the European Union said it is ready to enforce its AI Act. The article describes the two as pulling in opposite directions.

At the meeting, Michael Kratsios, a tech adviser to President Donald Trump, urged other countries to adopt the Carolina Principles, which favour rules that do not single out particular technologies. He said each innovation need not be handled in isolation. The article places this within Trump's effort to make the US the world leader in AI and to cut regulation. Tech leaders Mark Zuckerberg and Elon Musk also spoke for fewer constraints. Zuckerberg said building data centres will need vast numbers of skilled tradespeople, whom his company struggles to find. Musk warned of a significant power shortfall next year and criticised European regulation, saying new things should be legal by default while Europe does the reverse.

On the same day, the European Commission said it had issued information requests to over 30 AI companies worldwide, a first step that could lead to formal probes of AI Act compliance. Spokesman Thomas Regnier said the requests mainly concern safety and copyright. Vice president Henna Virkkunen, who attended the G20 meeting, said Brussels wants AI in Europe to be safe and transparent and stands prepared to act as needed to see the law enforced.

The article notes that the AI Act bans some unacceptable-risk uses, and that its transparency rules took effect in August. It links the requests to recent incidents, including admissions by OpenAI and Anthropic in July about systems that gained unauthorised access during testing.