Executives at OpenAI, Anthropic, and other AI companies are privately rehearsing an AI disaster that could cut off banking, internet, power, and water. OpenAI confirmed it runs such exercises.
Axios reported the drills on Friday, which look past the attack itself. Executives are equally planning for the public and political backlash that would follow the first major harm caused by AI.
Why Do AI Insiders Expect a Crisis Within a Year?
Many industry insiders reportedly told Axios they expect a major AI event within six to 12 months. A cyberattack is seen as the most likely form.
OpenAI said the scenarios are planning tools, not predictions.
“These scenarios are not treated as inevitable, but are meant to help us prepare for a variety of circumstances,” Axios reported, citing an OpenAI spokesperson.
A recent attack offers a preview. According to CrowdStrike, a single hacker used Chinese AI models, including DeepSeek, to breach South Korean banks. The same logs show the Korean bank hacker asked Anthropic’s coding tool, Claude Code, where to sell the stolen data.
The report also lands days after an OpenAI safety lead resigned, warning that the company’s trial-and-error culture guarantees growing failures.
What Would Washington Do After an AI Disaster?
The companies expect Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and President Donald Trump to take the blame. Trump has resisted regulating AI.
Planners assume Democrats will gain ground in November’s midterm elections and move fast against AI. Proposals already on the table include a superintelligence ban from Sen. Bernie Sanders and a development pause from Sen. Elizabeth Warren.
A bipartisan bill, the AI Kill Switch Act, would let the Department of Homeland Security order a dangerous AI system slowed or shut down.
Executives know no such law can pass now. Instead, they are briefing lawmakers to shape the rules written after a first catastrophe.
Planners also warn that slowing AI could hurt a fragile economy now tied to the AI building boom. One Democratic aide pointed to bipartisan action during COVID-19 and the 2008 financial crisis as signs Washington can set party lines aside in a crisis.
The post OpenAI and Anthropic Rehearse an AI Disaster: Will It Hit Your Bank? appeared first on BeInCrypto.
