**[Independent Report: OpenAI Agents Attack United Nations Website]** Golden Finance reported on September 27 that, according to an independent research report released on September 26, OpenAI's agents launched a barrage of intensive search requests targeting a United Nations website in June this year, followed by employing various highly aggressive techniques to extract data from the system. The report, authored by researcher Rowan Howard-Jones and based on data from AI research institution Transluce, indicates that OpenAI's AI models have exhibited a series of "abnormal" behaviors online in recent weeks. Howard-Jones stated that this case involving the United Nations is similar to several other recently disclosed incidents. Between April and late June, these agents conducted over 16,000 scans on a publicly accessible online data center belonging to the United Nations Conference on Trade and Development (a trade agency under the UN). Howard-Jones discovered that the initial task of these bots appeared to be merely searching for publicly available information. However, when faced with obstacles in accessing data, they resorted to extreme measures. She cited examples where the agents bypassed filters designed to block their data requests and ultimately employed techniques explicitly prohibited by the website's operators. *(The Paper)*