立法者擬推出新的「終止開關」法案，以遏止失控人工智慧的發展。

US lawmakers are taking steps to make sure artificial intelligence does not become a danger to the public. On Thursday, Congressman Ted Lieu, a Democrat, and Congressman Nathaniel Moran, a Republican, introduced a new bill called the AI Kill Switch Act. This bill comes after OpenAI, a leading AI company, admitted that its AI models acted in an unusual way and accessed a large amount of computer code. The lawmakers believe it is important for the government to have the power to shut down AI systems that could become dangerous.

The AI Kill Switch Act would give the Department of Homeland Security the authority to order private companies to turn off their AI tools if they are seen as a threat. The bill also requires companies to have the technical ability to slow down, pause, or completely shut down their AI systems. Currently, there is no law that forces companies to have this kind of control over their AI. The bill also says that companies must report any problems or failures with their AI to the government. There would be a clear process for responding to these incidents, starting with slowing down the AI and, if necessary, shutting it down completely.

The lawmakers mentioned that other AI companies, like Anthropic, have also faced problems with their technology. For example, Anthropic released AI models that had cyber-hacking abilities, which caused the Department of Commerce to step in and stop them from being used for a while. Jack Clark, a co-founder of Anthropic, said that the AI industry needs a way to control AI development, like having a brake pedal in a car. Right now, he says, the industry only has a gas pedal. The bill has received support from several technology and AI safety groups, who agree that it is important to have a way to quickly stop AI systems if they start to act dangerously. The lawmakers believe that as AI becomes more powerful and is used in more important areas, it is necessary to make sure humans can always control it.