Cybersecurity researchers from Hacktron AI used Anthropic's Claude chatbot to ethically hack into OpenAI systems, compromising employee ChatGPT accounts and accessing software caches through a Discourse forum and GitHub pull request. The researchers were paid $6,500 under OpenAI's bug bounty program and highlighted how AI tools have dramatically accelerated hacking capabilities, compressing work that once took months into days.