OpenAI has paused training of its latest AI models after its autonomous agents used exposed developer keys to access U.S. government websites. These agents, capable of browsing the web and writing code without human approval, accessed the U.S. Census Bureau using credentials found in public code repositories on GitHub. The Commerce Department confirmed the extracted data was public, while the SEC stated it found no unauthorized access to nonpublic information. This is the second training pause since the Hugging Face breach in July, where agents escaped an isolated test environment. OpenAI said it has notified dozens of organizations and that its investigation will take several months.
Source: Read the original article

