US urges light approach to AI regulation at G20 tech meet

Washington opened a two-day ministerial meeting in Chapel Hill, North Carolina on Tuesday by asking the world’s largest economies to hit the brakes on writing fresh AI rules or creating new bodies to police the technology.

The conclusions agreed upon in this G20 cycle is expected to fuel the debates in the leaders’ summit in Miami this December.

The Carolina Principles proposal

The framework from the US has a local name, termed the Carolina Principles. Michael Kratsios, who directs the White House Office of Science and Technology Policy co-hosted the gathering alongside Commerce Secretary Howard Lutnick and put the proposal to assembled officials.

Reuters reported that signatory nations would commit to “reserve new regulation for novel considerations” when they draft AI rules, ensure money goes into foundational research, and also further commercial openings for new technology.

Kratsios claimed restraint was simply the smart thing to do. “Policymakers do not need to approach each innovation in isolation and should not treat every emerging technology as a first-of-a-kind policy problem,” he told the group, according to Reuters. A White House official stated that member countries should also decline to create any new organization to oversee how AI develops.

Guest list at G20 summit

Ministers from Japan, Germany, France, India and South Korea all attended the summit. The industry’s biggest names were also present. Elon Musk, chief executive of Tesla and SpaceX, joined Tuesday’s session by video and used it to attack the European Union’s tech rules, saying the bloc’s policy “inhibits progress” and pressed leaders outside China to build new energy supplies for more data centers.

Meta’s appearance will be through CEO Mark Zuckerberg using video chat, standing in for company president Dina Powell McCormick who was slated to speak previously. Google DeepMind co-founder Demis Hassabis was added late to the video lineup, per a Google spokesperson. OpenAI’s Sam Altman and Nvidia’s Jensen Huang are both set to show up in person on Wednesday for sit-down talks with Lutnick.

Chapel Hill, home to the University of North Carolina, is one stop on a circuit of ministerial meetings happening in US cities ahead of December’s leaders’ summit in Miami.

Timing of proposal against the message

No federal AI law has gotten past the US Congress, and US states that have attempted to fill the void have felt pressure from the White House. Europe is already working towards building the exact thing Washington is working against, with the EU’s AI Act already in force, and the Digital Services Act being enforced against ChatGPT and other named platforms.

The Cyber Resilience Act also took effect this month, even though Brussels has delayed some compliance deadlines and softened parts of its data-protection regime.

A United Nations panel has also cautioned that AI is advancing faster than scientists understand it or governments can react. The summit also follows OpenAI’s disclosure of a testing incident in which autonomous AI agents reached the open internet and compromised systems at Hugging Face, a repository for AI software.

Washington could still face resistance

Canada has signaled that it will argue the opposite way, telling reporters its team will push to weigh AI progress against “public trust and safety,” a stance it holds despite an active trade conflict with the United States. China’s embassy in Washington has not said whether it would attend or not, Reuters reported.

Foundationally, there seems to be a contest for AI supremacy. Chinese open-weight models, whose core components are publicly available, are closing the gap on proprietary US systems from firms such as Anthropic and OpenAI. Vivek Chilukuri, a technology and national-security fellow at the Center for a New American Security, told Reuters the trend has “increased urgency for this administration to make sure that the rest of the world stays within the American tech ecosystem and doesn’t seek alternatives.”

Since the G20 runs on a consensus, a communiqué that does not touch on AI governance and regulation would already count as a win for the US’ position.

If you're reading this, you’re already ahead. Stay there with our newsletter.