344
Productivity & Workflow355
Automation & Workflow224
Software Development251
Marketing & Growth192
AI Infrastructure & MLOps174
Writing & Content Creation203
Data & Analytics141
Photography & Imaging156
Design & Creative170
Customer Support131
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
A new proposal would let Homeland Security order some AI systems to slow down or shut off in emergencies and would require companies to build the controls.
In short: US lawmakers are preparing a bill that would require some AI companies to build emergency controls that let the government order their systems to slow down or shut off.
Two members of the US House of Representatives, Ted Lieu (a Democrat from California) and Nathaniel Moran (a Republican from Texas), are expected to introduce a proposal called the “AI Kill Switch Act.” The idea is to make sure powerful AI systems can be stopped quickly in a serious emergency.
According to a Politico report cited by The Verge, the bill would give the Department of Homeland Security the ability to order an AI company to shut down its AI model, or “throttle” it, meaning force it to run in a limited, slower mode (like putting a car into limp mode). The bill says this could happen after consultation with the Secretary of Commerce and the Director of National Intelligence.
The proposal describes “loss-of-control” situations. Examples listed include incidents involving at least 10 deaths, more than $100 million in economic damage, or attempts by the AI model to hide or block its own shutdown controls.
The bill would also require major AI companies to build the tools needed to turn systems off, slow them down, and suspend user access. Companies would have to report safety incidents. Violating an emergency shutdown order could lead to penalties of up to $20 million per day.
Many people now rely on AI systems through apps and online services. This bill is meant to create a clear emergency brake when something goes seriously wrong, and to push companies to plan for worst-case scenarios instead of reacting after the fact.
Source: The Verge AI