The US is looking at a new bill called the AI Kill Switch Act to deal with the growing risks of advanced AI. The act was brought by Democratic Rep. Ted Lieu of California on July 23, 2026. The act would allow the government to step in and shut down AI systems if they become a serious threat to national security or public safety.
It requires certain AI developers to maintain the ability to slow or completely stop problematic systems. AI companies covered by the act must be able to terminate access or block a risky account. They must improve technical controls that can stop a model from running. However, in less-risky scenarios, the developer should reduce the computing power of a model.
The bill doesn’t cover all the AI companies. Instead, it would cover companies that earn at least $500 million in annual gross revenue from developing AI technologies. Thus, almost all the AI giants, including OpenAI, Google, Anthropic, Meta, and xAI, come under the law.
The government could use these powers only in major emergencies, such as large cyberattacks or other serious threats linked to AI. Reportedly, companies that violate general kill switch requirements will face a civil penalty of up to $2 million per day. Similarly, ignoring a DHS emergency order could raise that penalty to $20 million for every day if the company keeps violating the rules.
AI companies are working hard to build smarter and faster models. Every new launch brings better tools, stronger reasoning, and new features. At the same time, governments are trying to make sure these systems stay safe and do not create new risks.
The AI Kill Switch Act is still just a proposal. However, it shows how fast the conversation around AI is changing. Governments now view advanced AI as a technology that needs to be carefully overseen.
Also Read: China Bans AI Companion Personas as New AI Rules Take Effect
The future of AI will be driven by both innovation and safety. It's important to make better AI, but it might be even more important to keep it safe.