Security researchers at Hacktron reported using an AI model to find a flaw allowing them to take control of which company's public help forum?

Explanation: Researchers at security firm Hacktron said they used Claude to identify a security flaw in OpenAI's public help forum that allowed them to take control of the site, a disclosure that intensified debate over the dual-use potential of advanced AI models in offensive cybersecurity research. The case illustrates growing concern that AI assistance, even when used for legitimate vulnerability research, could lower barriers to discovering and exploiting weaknesses in widely used online platforms.