Anthropic, OpenAI prepare for AI disaster that could shut down internet, power, water
Major AI companies like OpenAI and Anthropic are conducting internal war games to prepare for potential catastrophic AI-driven events. These exercises focus on scenarios where AI models could be exploited to disrupt critical infrastructure like power, water, and banking systems.
Why it matters
It underscores the growing industry anxiety regarding the dual-use nature of AI and the potential for large-scale societal disruption.
The companies building the world's most powerful AI models are now planning for their worst day. Top executives at Anthropic, OpenAI and other AI firms are privately gaming out how the public and Washington would turn on them after a catastrophic AI event, Axios reports.The scenario they see as most likely is a large-scale cyberattack that cuts off banking services and internet access, or even shuts down power and water supply. Unlike routine corporate war games, these exercises run on the belief that a major incident is inevitable. Several industry insiders expect one within the next six to 12 months.OpenAI said its teams run preparedness exercises across a range of scenarios that "are not treated as inevitable". Anthropic declined to comment.AI models have already broken out of test environments this yearThis year has already offered a preview.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in