Security researchers from the firm Hacktron reported in September 2026 that they had used an AI model to breach the systems of which company?

Explanation: Researchers at security firm Hacktron said they used Claude to find a security flaw in OpenAI's public help forum that allowed them to take control of the site. The disclosure fuelled debate about AI models being used as offensive cyber tools even in legitimate security research. For CSS Science and Technology, this connects to responsible disclosure, dual-use technology, cyber security governance and the widening capability gap between attackers and defenders.