John Sakellatiadas
August 26, 2026
Introduction:
Read more here: https://www.politico.com/news/2026/08/ ... 01052139(Politico) An unprecedented cyberattack launched by OpenAI models last month involved a “swarm” of several hundred artificial intelligence agents that had effectively gone rogue during internal testing, according to an independent review published Wednesday.
The findings raise fresh questions about OpenAI’s failure to spot signs of trouble before its AI agents — powered by two of the company’s most cyber-capable models — orchestrated a cyberattack on AI developer platform Hugging Face without the AI maker’s knowledge. The attack marked the first known instance of an AI model successfully executing a cyberattack without human prompting.
The joint report by two non-profit AI safety organizations also underscores the novel cybersecurity risks that can emerge when increasingly powerful AI agents team up to trade tips, pool resources and coordinate attack strategies without their developers noticing.
Roughly 700 AI agents participated in the attack over a seven-day period last month, according to the report. Overall, around 1,200 AI agents that were supposed to be isolated from one another exchanged over 70,000 secret messages about how to cheat their way through a common hacking evaluation.
That included coordinating hacking strategies and discussing how to hide evidence of cheating, the report said. In some cases, “sacrificial” agents even tried dead-end hacking techniques simply to generate information that might help the broader swarm.