Evolving AI: OpenAI Just Got Hacked By an AI in 72 Hours!

A three-person security research team hacked OpenAI using AI — and the way they did it shows how quickly AI is changing cybersecurity. Hacktron AI researchers discovered a vulnerability chain involving OpenAI's community forum, Discourse, an image-processing dependency called libheif, and OpenAI's authentication system.

In this video, we break down the OpenAI security breach, how Claude Opus 4.8 and Claude Opus 5 were used during vulnerability research, and how the researchers eventually demonstrated access to an OpenAI employee's connected internal GitHub environment. Rather than stealing source code or private data, they proved access by having Codex create a harmless pull request before responsibly reporting the vulnerabilities.

The bigger story is AI-powered cybersecurity. Tasks that once required highly specialized security researchers and weeks of work can increasingly be accelerated by advanced AI models. We explore what this means for ethical hacking, AI security, vulnerability research, ChatGPT, OpenAI, Anthropic Claude, and the future of cybersecurity.





#OpenAI #Cybersecurity #AI #ClaudeAI #ChatGPT #EthicalHacking #ArtificialIntelligence