OpenAI ‘ethically hacked’ with help of Anthropic’s Claude chatbot
US cybersecurity researchers who conducted hack say ‘scope of what we could theoretically access was huge’
TL;DR
- Cybersecurity researchers successfully hacked into OpenAI employee accounts using AI chatbots like Anthropic's Claude and OpenAI's GPT-5.6 Sol.
- The hack involved compromising ChatGPT accounts and accessing software caches, with researchers noting the 'huge' theoretical scope of access.
- Hacktron AI, the research team, reported the hack to OpenAI under its bug bounty program and received a $6,500 payment.
- The incident underscores how AI tools are significantly reducing the time and resources needed for sophisticated cyberattacks.
- This follows previous security incidents at OpenAI, including an 'autonomous agent' hack on Hugging Face.
- OpenAI has acknowledged the vulnerabilities and stated they have been fixed.
- The event occurs amidst broader discussions about AI safety, with companies like Anthropic calling for a slowdown in AI development, a sentiment echoed by OpenAI and Google DeepMind, but rejected by Donald Trump.