Bipartisan lawmakers introduced the AI Kill Switch Act on Thursday, which would grant the Department of Homeland Security authority to shut down or throttle AI models deemed dangerous.
The legislation marks a significant shift in how Congress approaches AI regulation, moving from advisory frameworks toward direct intervention powers.
Under the proposed bill, DHS would gain authority to identify AI systems posing national security or public safety threats and take action to disable or reduce their capabilities. The agency could operate without waiting for traditional regulatory processes, enabling rapid response to emerging risks.
The bipartisan nature of the proposal signals growing consensus among lawmakers that AI systems require active oversight mechanisms. Both parties have expressed concerns about uncontrolled AI deployment, though they often emphasize different risks—national security threats versus potential harms to workers and consumers.
Details on the bill's specific triggers for intervention remain limited. Key questions include how DHS would define "dangerous," what appeals process developers could access, and how the agency would coordinate with existing regulators like the FTC and NIST.
The proposal comes amid accelerating AI development and deployment. Companies including OpenAI, Google, and Anthropic have scaled their systems rapidly over the past year, outpacing regulatory frameworks. Congressional interest in AI oversight has intensified accordingly.
Industry response will likely be mixed. Some companies welcome clear standards and government coordination on safety. Others may oppose what they view as excessive government power to restrict innovation and competition.
The bill faces an uncertain path through Congress. While bipartisan support is notable, implementation details could fracture that coalition. Debates over regulatory scope, due process, and innovation impact typically divide lawmakers along familiar lines.
The AI Kill Switch Act represents one of several legislative attempts to create binding AI governance rules. Other proposals focus on transparency requirements, liability frameworks, and safety testing mandates. The coming months will show whether Congress can maintain consensus as specific mechanisms come into focus.
OpenAI has begun rolling out GPT-6 Astra to Pro plan customers on its $100 and $200 monthly tiers. The release follows OpenAI's typical rollout pattern of prioritizing higher-tier subscribers before broader availability.
Current AI systems cannot yet independently design circuit boards, according to research from EEBench. The gap between AI capabilities and the complexity of PCB design remains significant.
Anthropic researchers have completed a formal mathematical proof of Fermat's Last Theorem, translating Andrew Wiles' decades-old proof into machine-verifiable code. The achievement marks a milestone in computational mathematics, ensuring the theorem's logical foundations are beyond dispute.
OpenAI has released GPT-6 Astra, its most advanced model to date, marking a significant stride toward artificial general intelligence. The company has implemented new safety guardrails due to the model's powerful cybersecurity capabilities.