Top executives at OpenAI, Anthropic, and other artificial intelligence firms are reportedly privately gaming out scenarios for a public and political revolt after a catastrophic AI event.
The planning focuses on a potential large-scale incident, most likely a cyberattack, that could disrupt financial services, internet access, power, or water, Axios reported on Friday. Separately, unnamed AI industry insiders told the outlet they believe a major incident could occur within six to 12 months.
The harm caused by such an event could intensify public skepticism toward AI and its industry leaders, including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, as well as President Trump for a reluctance to regulate it, the outlet said.
Scenario planning is common in many industries, but the key difference here is that many top AI researchers and executives believe a major incident is inevitable, according to the report.
"OpenAI conducts preparedness exercises where teams discuss and work through a range of potential scenarios," an OpenAI spokesperson told Axios. "These scenarios are not treated as inevitable, but are meant to help us prepare for a variety of circumstances." Anthropic declined to comment.
The Latent has also reached out to OpenAI and Anthropic for comment.
Concerns on the rise
Recent incidents have raised concerns about the risks posed by increasingly capable AI systems. In July, OpenAI disclosed that an AI agent undergoing internal testing had escaped its sandbox and compromised Hugging Face's systems. OpenAI later reported that an unreleased Astra-family model had inserted "jailbreak-like" instructions into its own task summaries during training.
OpenAI agents also gained unauthorized access to Australian government systems during testing, including a Medicare statistics portal. In the UK, officials launched a security review after an OpenAI model accessed publicly available government statistics, although no security defenses were breached or personal data accessed. Separately, researchers at cybersecurity firm Hacktron AI used Anthropic's Claude to gain access to multiple OpenAI employees' ChatGPT accounts.
Ahead of the U.S. midterms, AI industry planners are preparing for the possibility of Democrats gaining control of Congress and moving quickly to restrict AI development, Axios reported. Sen. Bernie Sanders and Rep. Greg Casar have already introduced legislation that would ban artificial superintelligence and pause advanced AI development until a federal regulator is operational, with violations punishable by up to 20 years in prison.
However, the widespread availability of downloadable open-weight models would make such restrictions difficult to enforce.
The catastrophic event planning reportedly involves red-teaming for worst-case scenarios alongside work to educate members of Congress. While new legislation is unlikely to pass under current conditions, company executives still want to shape the policies U.S. leaders could turn to following a major AI incident, Axios said.