Researchers Hack OpenAI With Claude

Security researchers used Claude to access an OpenAI employee's ChatGPT account.

•Balanced

AI security experts say they used Claude to hack ChatGPT

Hackers breached OpenAI, adding to fever pitch of security and safety concerns

Three Hackers Used Claude to Break Into OpenAI In Less Than 72 Hours

Articles

Sources

Coverage mix

What happened

The facts the coverage agrees on, drawn from reporting across the political spectrum. Where the sides dispute something, it's attributed.

Independent security researchers from Hacktron AI used Anthropic’s Claude chatbot to break into OpenAI employee ChatGPT accounts and reach OpenAI’s internal software environment. The entry path ran through Discourse, the third-party software behind OpenAI’s community forum, then through weaknesses in OpenAI employee validation and single sign-on. The researchers said the work was an authorized bug hunt under OpenAI’s reward program, and they reported the flaws instead of abusing them. OpenAI said it thanked them, narrowed community sign-in permissions, revoked affected tokens and sessions, and fixed the vulnerabilities. Hacktron said the chain took less than 72 hours. OpenAI paid a $6,500 bounty. The case now feeds wider concern about cyberattacks on leading AI firms.

Both sides

How left-leaning and right-leaning outlets told this story, in their own coverage. Outlet names link to the article we drew from.

From the left

Left-leaning outlets led with an AI-assisted breach story, focusing on Claude’s role in opening a path into OpenAI-linked systems. The Guardian framed the episode around OpenAI being ethically hacked with help from Anthropic’s chatbot. It also cited Hacktron saying AI tools had made the work far easier to carry out. The Verge emphasized restraint, noting the researchers stopped short of entering the code itself.

From the right

Right-leaning outlets treated the episode as a white-hat test that showed how hard complex AI systems are becoming to defend. Newsmax led with legal hackers using Anthropic’s Claude to penetrate ChatGPT-related systems. It stressed that the researchers disclosed the weaknesses rather than using them maliciously. Just the News placed the case inside a broader warning about AI agents and guardrails.

Coverage•8 articles — 2 left, 4 center, 2 right