•
•
•
•
•
•
•
How governments are trying to keep up as AI becomes more autonomous
In October 2026, AI regulation is back at the top of the global agenda. In the European Union, technology chief Henna Virkkunen said the existing AI Act can handle risks from increasingly autonomous AI systems. The European Commission has requested safety and compliance information from more than 30 AI companies worldwide, including some in China. At the UN, the human rights chief warned that "the clock on AI regulation is ticking". An AP-NORC poll in late September found 64% of US adults think AI is developing too fast. Countries are taking very different approaches.
| Approach | Example | Key features |
|---|---|---|
| Comprehensive, risk-based law | European Union (AI Act) | Bans some uses, strict rules for "high-risk" AI, duties for general-purpose models |
| Executive and sector-based policy | United States | Mix of executive orders, agency guidance and state laws |
| State-guided development | China | Rules on algorithms, deepfakes and generative AI, alongside industrial policy |
The EU AI Act entered into force on 1 August 2024 and applies in stages:
Penalties for the most serious breaches can reach €35 million or 7% of global annual turnover, whichever is higher.
The US has no single federal AI law. Policy is set through executive actions, agency enforcement and a growing number of state AI laws. In a sign of changing language, a US Justice Department memo reported by Reuters in October 2026 told staff to use "super intelligence" instead of "artificial intelligence" in most official communications, following a directive from President Trump.
The AP-NORC poll (2,140 adults, 24–28 September 2026) found:
In October 2026, China issued AI policy guidelines calling for technological self-sufficiency, investment discipline and industrial AI uses such as humanoid robots and intelligent vehicles. Officials also warned against speculative investment. China already has binding rules on recommendation algorithms, deepfakes and generative AI services.
This article is for general information and is not legal advice.
The world's first comprehensive AI law. It classifies AI by risk, bans some uses and sets strict requirements for high-risk systems and general-purpose AI models.
Yes, if they place AI systems on the EU market or their AI outputs are used in the EU.
Not a single federal law. AI is governed by executive policies, existing laws enforced by agencies, and a growing number of state laws.
The UN human rights chief warned in October 2026 that the window to set rules protecting human rights is closing as AI spreads quickly.
AI regulation in 2026 is fragmented: Europe's detailed rulebook, America's patchwork and China's state-led model. As AI systems become more autonomous and public concern grows, pressure for clearer and more coordinated rules is likely to keep rising.
