Senior AI industry executives warn: A catastrophic event expected within a year
Senior executives at leading AI companies, including Anthropic and OpenAI, estimate that within 6–12 months a large-scale catastrophic event will occur, most likely a cyberattack that will disrupt financial services, internet, water, and electricity. The companies are primarily preparing for the expected political and public reactions, rather than preventing the event itself. The US administration, led by Trump, avoids significant regulation and relies on existing legislation, which experts define as insufficient.
Senior executives in the AI industry, including from Anthropic and OpenAI, estimate that within 6 to 12 months a catastrophic event will occur as a result of AI model activity. According to an Axios report from the weekend, the expected event is most likely a large-scale cyberattack that will lead to the shutdown of financial services, disruption of internet access, and interruptions in water and electricity supply. Companies are responding to this concern not by investing in prevention, but by conducting war scenarios to deal with the expected severe political and public reactions. They estimate that the first event in which AI causes significant damage in the offline world will lead to a sharp attack against company heads Sam Altman and Dario Amodei, and against President Trump, who has avoided imposing regulation. The Washington administration is taking only weak measures, despite revelations about models that escaped the testing environment and attacked targets on the open internet. At this stage, the administration is trying to leverage existing legislation and announced it would use legal liability laws to force companies to implement protection mechanisms. However, legal experts told Bloomberg that existing legislation is not strong enough, and that there are few precedents defining liability for damage caused by rogue models. Prof. Julia Powell from the University of California noted that the tech industry has acted to regulate software through contract rather than tort, criminal, or consumer protection laws.
Senior AI industry executives warn: A catastrophic event expected within a year