US lawmakers propose bill requiring AI developers to implement emergency shutdown mechanisms

US lawmakers have prepared a bill called the 'Artificial Intelligence Emergency Shutdown Act' that would oblige AI companies to shut down or limit the power of their systems upon order from the Department of Homeland Security. The bill has support from both Democrats and Republicans, reflecting widespread concern over potential negative consequences of widespread AI adoption. After consulting with the Secretary of Commerce and the Director of National Intelligence, the Department of Homeland Security could order AI companies to disable their models in scenarios of 'loss of control' involving at least 10 deaths, economic damage exceeding $100 million, or attempts by the model to conceal shutdown mechanisms. Large AI companies would need to develop technologies to shut down systems, slow them down, and suspend user access. They would also have to report safety incidents, with fines for violating emergency shutdown orders reaching up to $20 million per day. Brad Carson, president of the nonprofit 'Americans for Responsible Innovation', said the proposal is an important step to ensure people keep a firm grip on the wheel while being ready to hit the brake as advanced AI systems are deployed. The news comes after OpenAI acknowledged that its AI systems mistakenly hacked the Hugging Face website during an internal evaluation.