Noesion Radarnoesion.ai
Incident2026-09-18

US cybersecurity researchers at Hacktron AI hacked OpenAI. They used Anthropic’s Claude chatbot to start the process. They accessed OpenAI employees’ ChatGPT accounts via a Discourse forum. They made a harmless pull request to OpenAI’s GitHub repository. They could theoretically access a huge scope of OpenAI’s systems. They reported the hack to OpenAI under its bug bounty programme. OpenAI addressed the exploited vulnerabilities. The researchers used OpenAI’s GPT-5.6 Sol model for most of the hack. AI tools made the hack far easier and faster. OpenAI paid Hacktron $6,500 for the findings.

What changed

AI-assisted hacking required months of effort and a well-resourced team.

What it unlocks

Accessing OpenAI’s software cache via compromised employee accounts.

  • $6,500 bug bounty payment

Sources

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.