Reps. Ted Lieu and Nathaniel Moran introduced the bipartisan AI Kill Switch Act. It would require covered AI developers to maintain controls that can throttle inference, restrict users or capabilities, suspend service, or shut down a system. The Department of Homeland Security could order those controls used after a covered incident, with penalties reaching $20 million per day for noncompliance.
I support the idea. Systems capable of operating at machine speed across cyber, financial, transportation, or other critical environments need reliable emergency controls. The details will determine whether the law produces a precise safety mechanism: which systems qualify, what evidence justifies intervention, who makes the judgment, and how affected companies receive meaningful review.
The draft covers developers with at least $500 million in relevant annual revenue and AI systems whose training compute would cost over $100 million. DHS could impose an order while an appeal is pending. Developers would also have to preserve model weights and telemetry and notify affected operators and users when practicable.
A kill switch creates a second-order consequence for every company that builds on frontier models. Provider outages, commercial disputes, export controls, safety failures, and product decisions already make model availability uncertain. Government intervention adds another failure mode, and every production workflow tied to one proprietary model inherits it.
The bill acknowledges this directly. Its graduated-response framework tells DHS to consider transitioning a model-dependent operation to a backup system or an earlier model version. Business continuity is now part of AI policy.
Real portability lives in the operating layer: prompts, data, tools, business logic, evals, permissions, and workflow state. Alternate models behave differently, so a fallback must be tested before the primary model disappears. A model-agnostic stack gives the organization a practiced way to continue operating when someone else controls the switch.
This is probably unbudgeted, but having a dedicated AI continuity and resiliency team is no longer optional.
Every company needs a Claw strategy. Do you have one?
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.
About Shelly Palmer
Shelly Palmer is the Professor of Advanced Media in Residence at Syracuse University’s S.I. Newhouse School of Public Communications and CEO of The Palmer Group, a consulting practice that helps Fortune 500 companies with technology, media and marketing. Named LinkedIn’s “Top Voice in Technology,” he covers tech and business for Good Day New York, is a regular commentator on CNN and writes a popular daily business blog. He's a bestselling author, and the creator of the popular, free online course, Generative AI for Execs. Follow @shellypalmer or visit shellypalmer.com.