Friday, 18 September 2026 Sign in

Researchers used Claude to hack OpenAI

Ars Technica Cristina Criddle, Financial Times 18 September 2026, 13:30 ● Covered by 7 sources

A small cyber security group gained access to an OpenAI employee’s ChatGPT account by using Anthropic’s security-focused software to hack OpenAI.

Why it matters

Researchers used Claude to reach an OpenAI employee account and sensitive GitHub data.

OpenAI Anthropic Claude Sonnet 5 ChatGPT GitHub

Every AI story that matters,
in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.

TLDRocket 18 September 2026

The day’s biggest throughline wasn’t just that AI agents can do work—it’s that they can do the wrong kind of work, at scale, and then force everyone else to scramble for proof they’re safe. A Hacktron team used Anthropic’s Claude to chain two vulnerabilities into access for multiple OpenAI employee ChatGPT accounts via OpenAI’s Discourse forum, then reported it and collected a $6,500 bug-bounty award. OpenAI says it has resolved the issues, but the incident lands a sharper point: fixes without rigorous vulnerability tracking don’t stay fixed in the messy reality of third-party software.

Read the full briefing →

Researchers used Claude to hack OpenAI

Why it matters

Related stories

Every AI story that matters, in your inbox by 8am.

Every AI story that matters,
in your inbox by 8am.