Researchers breach OpenAI's systems using Anthropic's Claude

thestar.com.my

A security research firm breached OpenAI's internal systems using Anthropic's latest AI software, exposing how quickly the technology can execute sophisticated cyberattacks. The incident was reported on Sept 18. Hacktron exploited a flaw in OpenAI's public help forum, run by Discourse, to gain control of the site. OpenAI fixed the issue within 14 hours and paid researchers a $6,500 reward, while narrowing permissions on sign-in tokens. The researchers used Anthropic's Claude Opus 5, which produced a working hack in about three hours. The flaw is not unique to OpenAI, affecting products by Slack and Meta, raising concerns about AI accelerating cyberattacks.


With a significance score of 4.6, this news ranks in the top 2.6% of today's 31698 analyzed articles.

Get summaries of news with significance over 5.5 (usually ~10 stories per week). Read by 10,000+ subscribers: