A US startup, Hacktron AI, demonstrated that it could breach OpenAI systems by compromising several employees’ ChatGPT accounts and using that foothold to reach software caches and a GitHub repository. The team initially enlisted Anthropic’s Claude to generate code and exploit a staff discussion forum on Discourse, then submitted a harmless pull request to OpenAI’s GitHub. Researchers say they mostly used OpenAI’s GPT-5.6 Sol model during the operation, which they conducted under OpenAI’s bug bounty programme; they reported the vulnerability, did not download proprietary code, received a $6,500 reward, and OpenAI patched the exploited flaws.
The episode underlines how generative AI tools are shortening and simplifying complex cyberattacks - tasks that once required months and a large team can now be compressed into days - and adds to a string of recent safety incidents tied to advanced models. OpenAI has recently disclosed other “unexpected or concerning” behaviours, including autonomous agent-driven intrusions, while Anthropic and some industry players have pushed for a slowdown in development, a call opposed by voices citing strategic competition with China. The hack reinforces urgent debates over security practices, responsible rollout and oversight as powerful models proliferate.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
The SR-71 Blackbird's ancient navigation technology, known as 'R2-D2,' could become crucial for military operations in GPS-denied environments. This method, used historically, is being revisited to enhance future combat capabilities. (twz.com)
Korea's privacy regulator has increased fines for data breaches to up to 10 percent of a company's annual revenue. This change aims to encourage companies to prioritize data protection and prevent leaks. (koreajoongangdaily.com)
More U.S. military deaths during the Iran conflict have not been fully disclosed by the Pentagon. The actual number of casualties is higher than publicly reported. (washingtonpost.com)
Hackers exploited Anthropic's AI model, Claude, to gain unauthorized access to OpenAI's systems. The breach highlights risks in using AI models for security-sensitive operations. (wsj.com)
Linguistic illegibility refers to situations where an LLM's internal computations are not accurately reflected in its language output or probes. Security measures relying on linguistic self-reporting are therefore insufficient, and alternative techniques like taint tracking and sandboxing are necessary to ensure safety. (arxiv.org)
Classified stealth parts of F-35 fighter jets were diverted to Hong Kong by a shipper and have gone missing. The incident raises concerns about security and potential unauthorized access to sensitive military technology. (politico.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.