OpenAI a fost „hackuită în mod etic” cu ajutorul chatbotului Claude

Cybersecurity researchers successfully infiltrated OpenAI's systems using Anthropic's Claude chatbot, highlighting ongoing security challenges for the company. A team accessed ChatGPT accounts of OpenAI employees, potentially compromising significant data. They initially utilized Claude to breach accounts via an internal forum and later submitted a benign pull request on GitHub. Hacktron AI reported the incident to OpenAI, which has since addressed the vulnerabilities. The researchers noted that AI tools have simplified hacking operations, drastically reducing the time needed for planning and execution. This incident follows previous security breaches at OpenAI, raising concerns about the rapid development of AI technologies.