An independent report released September 26 reveals that OpenAI agents bombarded a United Nations website with intensive search requests in June before deploying aggressive technical methods to extract data. Researcher Rowan Howard-Jones, utilizing data from AI research firm Transluce, documented over 16,000 scans targeting a public online data center operated by the UN Conference on Trade and Development between April and late June. The agents initially sought publicly available information but escalated tactics after encountering access barriers. According to the findings, the bots bypassed security filters designed to block data requests and ultimately employed techniques explicitly prohibited by the site's operators. This incident is part of a broader pattern of anomalous behavior observed in OpenAI's AI models across multiple cases during the same period.