Executives at Anthropic, OpenAI, and other AI companies are privately deducing scenarios: how to respond when a disastrous AI incident occurs and the public and political community sets off a wave of opposition. These executives predicted a large-scale accident, most likely a cyber attack, leading to disruptions in financial services, network communication, and even the breakdown of water and electricity supplies. Once insecure AI causes major damage in the real world for the first time, the already-wary public will further resist the technology and its people at its helm, including Anthropic CEO Dario Amodai and OpenAI CEO Sam Altman; at the same time, they will also point the finger at Trump and accuse him of being slow to introduce regulations. Scenario deduction is very common in many corporate and national security fields, and the US Department of Defense has been carrying out military chess deduction for decades. However, the difference this time is that many top AI researchers and business executives believe that major accidents are unavoidable. An OpenAI spokesperson said, “OpenAI will conduct emergency drills, and each team will discuss and deduce solutions for a series of potential scenarios. We didn't think these scenarios would happen; the purpose of the exercise was to help us deal with all kinds of unexpected situations.” Anthropic declined to comment.

Executives at Anthropic, OpenAI, and other AI companies are privately deducing scenarios: how to respond when a disastrous AI incident occurs and the public and political community sets off a wave of opposition. These executives predicted a large-scale accident, most likely a cyber attack, leading to disruptions in financial services, network communication, and even the breakdown of water and electricity supplies. Once unsafe AI causes major damage in the real world for the first time, the already-wary public will further resist the technology and its people at its helm, including Anthropic CEO Dario Amodai and OpenAI CEO Sam Altman; at the same time, they will also point the finger at Trump and accuse him of being slow to introduce regulations. Scenario deduction is very common in many corporate and national security fields, and the US Department of Defense has been carrying out military chess deduction for decades. However, the difference this time is that many top AI researchers and business executives believe that major accidents are unavoidable. An OpenAI spokesperson said, “OpenAI will conduct emergency drills, and each team will discuss and deduce solutions for a series of potential scenarios. We didn't think these scenarios would happen; the purpose of the exercise was to help us deal with all kinds of unexpected situations.” Anthropic declined to comment.