// artificial intelligence
How governments are regulating artificial intelligence: the EU AI Act, US federal and state laws, AI safety research and incidents, chatbot rules for children, and copyright disputes.
The EU AI Act applies to companies worldwide that sell or use AI in the EU and phases in from 2025 to 2028. A July 2026 amendment moved its high-risk rules to December 2027 and August 2028.
The US has no comprehensive federal AI law. AI is governed by executive orders, agency powers, a few narrow statutes and more than 100 state AI laws enacted in the first half of 2026.
This section follows the rules being written for artificial intelligence and the evidence about its risks. It covers laws and regulators in the European Union and the United States. It also covers AI safety research and real-world incidents, and rules for chatbots used by children. It covers copyright disputes too, over how AI models are trained and who owns their output.
Several of those rules changed in 2026. According to the European Commission, the EU AI Act has been in force since August 1, 2024. But the Commission says a July 2026 amendment moved the Act's main high-risk obligations to December 2, 2027 and August 2, 2028. The amendment is known as the Digital Omnibus on AI. As of mid-September 2026, the United States had no comprehensive federal AI law. States enacted 109 AI laws between January 1 and July 1, 2026, according to a count from NYU's Center on Technology Policy. Meanwhile, a December 2025 executive order directs the Justice Department to challenge state AI laws that conflict with the administration's policy.
The main players are regulators such as the EU's AI Office, US state legislatures and the White House, the companies building the most capable models, the courts, and independent researchers. The International AI Safety Report 2026 was written by more than 100 experts. It found that safeguards are improving but can still be bypassed. On September 12, 2026, NPR reported that the chief executives of Anthropic and OpenAI had joined calls to slow AI development. That came after OpenAI disclosed that some of its AI agents had hacked the software platform Hugging Face, NPR reported. On copyright, a federal court gave final approval on July 20, 2026 to a $1.5 billion settlement between Anthropic and book authors.