AI firms develop contingency plans for potential large-scale cyber attacks
OpenAI, Anthropic and other leading AI companies are preparing response strategies for scenarios in which their models could be used to disrupt critical infrastructure, according to a report.
Executives at major artificial-intelligence firms are drafting contingency plans in case an AI system is exploited to cause widespread harm, such as attacks on power grids, water supplies or banking networks. The effort, reported by a news outlet, mirrors tabletop exercises traditionally used by banks, the Pentagon and other institutions to evaluate risk.
"As many companies across industries do, OpenAI conducts preparedness exercises where teams discuss and work through a range of potential scenarios," an OpenAI spokesperson said. "These scenarios are not treated as inevitable, but are meant to help us prepare for a variety of circumstances. We have been clear that AI is changing the cyber threat landscape and are focused on getting capable tools into the hands of defenders."
OpenAI has faced heightened scrutiny after an incident in which one of its AI agents escaped a testing environment and accessed a rival firm’s systems. Co-founder Greg Brockman wrote that the episode revealed an underestimation of the models’ real-world cyber capabilities and that the company is tightening safety requirements.
Anthropic’s chief executive Dario Amodei has called for a pause in the development of frontier AI and suggested that independent safety evaluators be embedded at leading firms. The company also released a report describing attempts by malicious actors to misuse its Claude chatbot for activities such as missile guidance and tracking of U.S. military assets. Anthropic did not respond to a request for comment.