The Carolina Principles for Emerging Technologies are a set of principles on the regulation of AI and other emerging technologies that the United States promoted at the G20 innovation ministerial it co-hosted in Chapel Hill, North Carolina, on September 1–2, 2026. White House science and technology adviser Michael Kratsios urged delegates on September 1 to adopt them, describing an agreement under which signatories would avoid writing entirely new AI regulations and would write rules only for "novel" situations involving the technology (Source: reuters.com; insideaipolicy.com).
Adoption
Kratsios told reporters that the Chinese government, represented by Technology and Science Minister Yin Hejun, had signed, but he did not provide a copy of the document (Source: reuters.com). The G20 leaders' summit is scheduled for Miami in December 2026 (Source: reuters.com). The full text and the list of signatories were not available in the coverage cited here.
Content as described
Beyond Kratsios's summary, the principles have been characterized in secondary accounts. A September 2026 Center for American Progress report states that the United States "secured G20 support from the Chinese delegation" for the principles and describes them as encouraging governments "to use regulatory sandboxes and experimental exemptions from certain regulatory requirements to accelerate the development and commercialization of emerging technologies"; its citation for that description is a September 3, 2026 article on the Winzheng site rather than the text itself. The report presents the principles as the international extension of the administration's deregulatory AI agenda, alongside America's AI Action Plan and Sen. Ted Cruz's SANDBOX Act (The SANDBOX Act Uses AI Policy as a Trojan Horse for Unfettered Deregulation (Center for American Progress, September 2026)).
Reactions at the ministerial
Other speakers at the ministerial took differing positions on regulation. Google DeepMind's Demis Hassabis called on the G20 to set up safety tests for AI systems; Meta's Mark Zuckerberg said countries should not restrict open-weight models; and Elon Musk said EU policy "inhibits progress" (Source: reuters.com). Anthropic co-founder Tom Brown, OpenAI's Sam Altman and Nvidia's Jensen Huang were scheduled to address the meeting on September 2 (Source: reuters.com).
Relationships
- related: Michael Kratsios — promoted the principles at the ministerial
- related: America's AI Action Plan — domestic policy the principles parallel
- related: AI Regulatory Sandbox, SANDBOX Act (S. 2750) — sandbox and exemption approach the principles are described as encouraging
- related: The SANDBOX Act Uses AI Policy as a Trojan Horse for Unfettered Deregulation (Center for American Progress, September 2026) — secondary characterization of the principles