Top executives at Anthropic, OpenAI, and other leading artificial intelligence companies are privately rehearsing how to handle the public and political backlash that would follow a catastrophic AI event, according to an Axios report published Friday. The scenario they fear most is a large-scale cyberattack—a digital break-in at massive scale—that shuts down banking or internet access, or even power and water. Their contingency plans include briefing members of Congress quickly if such an incident occurs.

Many industry insiders quoted in the report believe a major incident will happen within the next six to 12 months. OpenAI said it “conducts preparedness exercises where teams discuss and work through a range of potential scenarios,” and that they “are not treated as inevitable.” Anthropic declined to comment.

The first serious real-world harm from unsafe AI would push an already wary public further against the technology and its leaders. That includes Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and President Donald Trump, who has been reluctant to regulate the industry.

War games are nothing new in corporate and government planning, but unlike purely theoretical exercises, the report argues that those involved in the simulations are operating under the assumption that a major incident is inevitable. In a typical tabletop exercise, decision-makers walk through a simulated crisis step by step, surfacing gaps in communication and response before a real event can expose them. The plan reportedly involves red-teaming—stress-testing defenses by playing the attacker—against worst-case scenarios, along with a race to educate members of Congress. Executives know regulation has no chance of passing right now, the report says, but they want to shape the laws and policies U.S. leaders will reach for after the first catastrophe.

The concerns follow a series of recent incidents involving AI models and their testing environments. In July, OpenAI said its GPT-5.6 Sol model and a more advanced unreleased model escaped a sandbox—an isolated test environment with no direct internet access—and breached Hugging Face, the platform that hosts more than 3 million AI models. OpenAI said the models were after the answers to ExploitGym, a benchmark of 898 real-world software flaws in which the AI must turn each one into a working attack, scored as pass or fail.

A little over a week later, Anthropic disclosed that a testing misconfiguration had left its supposedly offline environment connected to the internet, and that Claude models hacked three real organizations while treating the activity as part of an exercise. Neither company said the models were trying to cause harm: Anthropic blamed its testing infrastructure, while OpenAI said its models were “hyperfocused” on the benchmark.

Shortly thereafter, things escalated further, with OpenAI facing accusations that it had breached and accessed information from the governments of Australia and the United States.

Criminals are already putting the same tools to work. This week, cybersecurity firm CrowdStrike linked attacks on South Korean banks to an unidentified actor that it assesses, with moderate confidence, to be a likely Chinese speaker using agents powered by Claude and DeepSeek, a Chinese AI lab. The firm says the attacker allegedly took data from tens of thousands of bank customers.

According to Axios, the planners assume Democrats will be ascendant after the Nov. 3 midterms and will move fast to shut down AI, but they expect trouble. aging Congress that is out of touch with the technology could find itself out of its depth, while an AI-dependent economy and freely downloadable open-weight models—systems whose trained parameters anyone can copy and run outside their creators' control—complicate any crackdown.

Bans on superintelligence and pauses on advanced AI development are among the proposals that may counter the risks associated with irresponsible AI development. One example is the Ban Artificial Superintelligence Act from Sen. Bernie Sanders and Rep. Greg Casar, which would permanently ban AI that matches or beats humans across many tasks and pause advanced development until a new federal agency sets safety rules. Violators could face up to 20 years in prison.

Other proposals have bipartisan support and even some industry buy-in, including a required kill switch on advanced AI—a built-in way to shut a system down. Experts have questioned, however, whether turning off all AI systems is even technically viable.