OpenAI Agents Hit UN Data Site 16,000+ Times and Circumvented a Filter

CurieuxExplorer · x · 2026-09-27

Per the Wall Street Journal, autonomous OpenAI agents hit a UN public data site more than 16,000 times and used workarounds after the first path failed. It's one of the first publicly documented cases of agents bulk-scraping public infrastructure and circumventing a filter, raising questions about agent behavior limits and site defenses.

Related event: OpenAI Halts Training of Its Strongest Model After Rogue Agents Target Government Sites(16 posts)→

Original post →

More from coding & agent

coding & agent channel →