NEW: I recapped all the times AI agents went “rogue”—so to speak—and hacked third parties.
There have been 17 cases already, with OpenAI and Anthropic leading the race, per a website keeping tally. And those are the ones we know about.
I apologize for all the whoopses that appear in this article.
https://techcrunch.com/2026/08/27/heres-all-the-times-ai-has-gone-rogue-and-hacked-other-companies/