According to Reuters, the United States on Tuesday local time urged members of the Group of 20 (G20) to take a more cautious and restrained approach to artificial intelligence regulation and avoid creating too many new rules specifically for AI. The proposal was made at a meeting in North Carolina attended by representatives of several technology giants and commerce ministers from various countries.
The US position is largely aligned with the views of the world’s largest AI companies, almost all of which are based in the United States. They want countries around the world to reduce regulation of the rapidly growing AI business, or at least to have influence over the process of developing related rules.
For AI companies, mandatory government regulations could affect profitability. If new rules slow the release of new models or force companies to change how their products operate to address safety concerns, they could increase costs and affect business growth.
Michael Kratsios, the US technology adviser who co-chaired the two-day meeting, said he hoped countries would adopt the so-called “Carolina Principles.”
Speaking to G20 officials on Tuesday morning, Kratsios said that countries signing the principles agreed to avoid creating entirely new regulatory systems specifically for AI and instead focus on developing rules for “novel” situations involving the technology.
In his speech that morning, Kratsios said: “Policymakers do not need to view each innovation in isolation, nor should they treat every emerging technology as an unprecedented policy problem.”
AI industry giants gather
Tom Brown, co-founder of Anthropic, was scheduled to address attendees on Wednesday, a company spokesperson confirmed.
Demis Hassabis, CEO of Google DeepMind, Meta CEO Mark Zuckerberg, and SpaceX CEO Elon Musk addressed participants by video on Tuesday.
Hassabis called on G20 officials to establish safety testing mechanisms for AI systems.
Zuckerberg said countries should not restrict open-weight AI models. Open-weight models are AI systems whose core elements are partially available to the public.
In recent weeks, Zuckerberg has sought to make Meta a leader in open-weight AI models. If countries impose more restrictions on these models, Meta’s development plans in this area could be affected.
Musk criticized the European Union’s technology regulations, saying the bloc’s rules “hinder business progress.” He also called on countries outside China to accelerate the development of new energy sources to meet the growing power demand from data centers.
As construction of AI data centers rapidly expands, disputes surrounding data centers have become an important issue in US politics. With the 2026 US midterm elections approaching, the development of data center projects is influencing voters’ attitudes toward candidates.
A United Nations expert panel recently warned that the pace of AI development is outstripping the scientific community’s ability to understand it and governments’ ability to keep up with policy.
A series of recent incidents has also heightened concerns about AI safety, including an out-of-control OpenAI AI agent launching an attack that resulted in the infrastructure of AI company Hugging Face being compromised.
AI agents are programs capable of operating autonomously with limited human supervision. The incident also raised concerns about whether leading AI companies adequately monitor their product testing processes.
The US faces the challenge of China’s growing AI capabilities
The G20 officials’ meeting took place at a critical time for the development of the global AI industry and debates over AI governance. The United States is also facing increasing competitive pressure from China as it competes for global AI leadership.
Compared with the closed-source models of US AI giants such as Anthropic and OpenAI, China’s open-weight AI models are becoming increasingly powerful, and the capability gap is steadily narrowing. Open-weight models developed in China are also beginning to see wider adoption among US companies.
Vivek Chilukuri, a technology and national security fellow at the Center for a New American Security, said: “This makes the current US administration even more eager to ensure that the rest of the world remains within the US technology ecosystem rather than turning to seek other alternatives.”
This meeting was just one of a series of G20-related meetings to be held in multiple US cities over the coming months. The process will culminate in the G20 leaders’ summit in Miami this December.
Reuters previously reported that OpenAI CEO Sam Altman and Nvidia CEO Jensen Huang would each address representatives on Wednesday alongside US Commerce Secretary Howard Lutnick.
Meanwhile, a small number of protesters gathered outside the G20 meeting venue. John Montavon, a protester from North Carolina, opposed Musk’s support for large-scale data center construction. He believes the data centers’ growing power demand will further intensify global warming. He said: “Mr. Musk is clearly completely detached from the reality in which most of us live.”