OpenAI says it stopped a campaign to steal its models' reasoning, but the trick still worked on Azure

OpenAI reported that it blocked a coordinated effort involving over 15,000 accounts aimed at extracting the proprietary reasoning processes of its models, attributing part of the campaign to individuals linked to Moonshot AI. While OpenAI claims to have neutralized the threat, researchers discovered that the extraction technique remained functional on Microsoft Azure for several additional weeks. This incident highlights ongoing vulnerabilities in large language models and raises concerns about the security of intellectual property when hosted on third-party cloud infrastructure.