Agora issue guide
Should AI be regulated? The debate explained
By the Agora desk · Published · How we count player splits
There is no single U.S. law governing artificial intelligence. State laws, older rules on fraud, discrimination and privacy, and voluntary promises by the companies fill the gap, while the European Union has passed a broad AI Act. Supporters of new rules point to risks from deepfakes to losing control of powerful systems; opponents say rules written now would slow innovation and hand the lead to China.
Agora players, live
Do you support building more AI data centers to win the AI race, even if local communities bear the costs?
Yes 27%No 73%
104 players · Trump says people against data-centers are "backwards and poor", 1 September 2026 · Cite or embed
Latest Agora stories on this issue
What "regulating AI" can mean
The debate covers very different things. Some rules target uses: deepfakes, AI in hiring, lending or policing, chatbots talking to children. Others target the most powerful "frontier" models themselves, asking their makers to test for dangerous abilities, report incidents and publish safety plans. A third set covers the industry around AI, from chip exports to the power and water that data centres use.
Some of the starkest warnings come from inside the industry. In 2023 hundreds of AI scientists and executives signed a one-line statement that reducing the risk of extinction from AI should be a global priority. Agora asked players whether the U.S. should prioritise regulating frontier AI even if China doesn't and whether AI should be allowed to act on its own.
Who makes the rules now
In the U.S. most AI-specific laws so far are state laws. California passed a law in 2025 requiring the largest AI developers to publish safety frameworks and report serious safety incidents. Federal agencies apply existing laws on fraud, discrimination and consumer protection to AI, and the leading companies have made voluntary safety commitments. Agora asked players whether AI companies should be allowed to police themselves.
The European Union's AI Act, in force since August 2024, sorts AI by risk: some uses are banned, "high-risk" uses such as hiring and credit scoring face strict duties, and makers of general-purpose models must meet transparency rules. Its obligations phase in over several years and apply to American companies selling in Europe.
Washington vs the states, and the race with China
The sharpest U.S. fight is over who gets to regulate. In July 2025 the Senate voted 99–1 to strip a proposed 10-year freeze on state AI laws from a budget bill. In December 2025 President Trump signed an executive order setting up a Justice Department task force to challenge state AI laws, and the White House has asked Congress for a national standard that would override them. An executive order cannot by itself cancel a state law, so the question now sits with Congress and the courts.
Behind it is the race with China. Those who want light rules say America must build faster, including the data centres AI runs on; others say safety has to come first. Agora players argued whether regulation should come before winning the AI race against China and whether to build more data centres even if local communities bear the costs.
The case for regulating AI
- Powerful systems are released faster than anyone can test them, and their own makers warn of serious risks.
- Deepfakes, scams and biased automated decisions already cause harm that older laws weren't written for.
- Companies racing for market share have reasons to cut corners, and voluntary promises can't be enforced.
- Clear rules build the public trust the industry needs to grow.
The case against heavy regulation now
- Rules written before the technology is understood could lock in today's giants and shut out start-ups.
- Slowing American labs hands the lead to China, which won't slow down.
- Fifty different state laws make compliance costly; one light national standard is better.
- Existing laws on fraud, discrimination and product safety already apply to AI.
Common questions
Is AI regulated in the US?
Partly. There is no single federal AI law. Existing laws on fraud, discrimination and privacy apply, several states have passed AI-specific laws, and the federal government is pushing for one national standard that would override them.
Why does the White House want to stop states regulating AI?
It argues that 50 different state laws would be costly and slow U.S. companies in the race with China. Opponents, including state officials from both parties, say states should be free to protect their residents while Congress hasn't acted.
What is the EU AI Act?
A European law, in force since 2024, that bans some uses of AI, sets strict rules for high-risk uses like hiring and credit, and requires makers of general-purpose models to meet transparency duties. It covers any company selling AI in the EU, including American ones.
Key facts and sources
- The EU AI Act entered into force on 1 August 2024, with obligations phased in over the following years. (European Commission)
- The Senate voted 99–1 to strip a 10-year preemption of state AI laws from the 2025 budget bill. (Ropes & Gray, March 2026)
- Executive Order 14365 (11 December 2025) created a Justice Department AI Litigation Task Force to challenge state AI laws. (The White House)
- California's SB 53 (2025) requires large frontier AI developers to publish safety frameworks and report critical safety incidents. (Office of the Governor of California, 29 September 2025)
- In 2023 AI researchers and executives signed a statement that mitigating the risk of extinction from AI should be a global priority. (Center for AI Safety)
What Agora players think
The side players picked before playing each story. Live from Agora games, not a scientific poll.
Should AI companies be allowed to police themselves?
Yes 33%No 67%
39 players · Trump gets AI giants to agree to “self-police” AI development, 30 September 2026 · Cite or embed
Would you invest in Anthropic at its IPO?
Yes 40%No 60%
42 players · Anthropic IPO warns AI could pose an ‘existential risk to humanity’, 29 September 2026 · Cite or embed
Should AI be allowed to act on its own?
Yes 17%No 83%
35 players · OpenAI says its models may have interfered with government sites, 26 September 2026 · Cite or embed
Should the US prioritise regulating frontier AI, even if China doesn’t?
Yes 78%No 22%
78 players · Anthropic researcher quits and says AI could "Kill us all", 10 September 2026 · Cite or embed
Should AI regulation be prioritized over winning the AI race against China?
Yes 78%No 22%
92 players · Open AI agents hijack a German website, 6 September 2026 · Cite or embed
Do you support building more AI data centers to win the AI race, even if local communities bear the costs?
Yes 27%No 73%
104 players · Trump says people against data-centers are "backwards and poor", 1 September 2026 · Cite or embed
Every Agora story on this issue
Written by the Agora team from the sources above. Player numbers update as people play.
