A new bill in Congress would give the Department of Homeland Security the power to throttle or shut down certain artificial intelligence systems. The proposed 'AI Kill Switch Act' also requires that AI systems come with built-in shutdown controls. The legislation arrives as prediction markets show heavy betting on Anthropic to become the top AI model by July.
What the AI Kill Switch Act Would Do
The bill, introduced by lawmakers, targets what they see as a gap in current oversight. Under the act, the DHS Secretary could order specific AI systems to be slowed or completely disabled if they pose a risk. Companies would also have to include a built-in mechanism for such a shutdown. The proposal doesn't name any particular AI system, but it comes amid growing concerns about rapid AI development and potential misuse.
Supporters argue the measure is a necessary safety net. Critics worry it could stifle innovation or be used arbitrarily. The bill is still in early stages, with no committee hearings scheduled yet.
Polymarket's 98% Prediction
Meanwhile, Polymarket, a decentralized prediction market, gives Anthropic a 98% chance of being the top AI model by July. That's a strong bet on the company behind the Claude chatbot. Polymarket users are putting real money on that outcome. The prediction doesn't specify which metric defines 'top' — it could be user adoption, benchmark scores, or something else.
Anthropic has positioned itself as a safety-focused AI lab. The company has called for regulation before. But the 98% figure suggests traders see it as the clear frontrunner in the near term, at least.
The two stories — the kill switch bill and the Polymarket bet — highlight a tension. Lawmakers are moving to impose controls on AI, while markets are betting on a specific company to dominate. Whether the bill would affect Anthropic's trajectory is unclear. The prediction market's deadline is July, giving a concrete timeline to watch.
