Researchers used Claude to help breach OpenAI employee accounts

Bug-bounty researchers chained multiple vulnerabilities to gain access to OpenAI employee ChatGPT accounts and an internal code repository. The attack path was disclosed and remediated after the researchers reported their findings.

The incident demonstrates how AI tools can accelerate conventional attack techniques, allowing adversaries to move through systems more quickly. It also highlights the ongoing importance of third-party identity systems, which remain a critical component in securing enterprise environments.

Source: https://www.nbcchicago.com/