🇺🇸🇺🇳⚡️ — OpenAI agents made more than 16,000 requests to a United Nations data platform and used aggressive methods to bypass restrictions designed to block them, according to an independent research report cited by WSJ.

The activity targeted a publicly accessible UN Trade and Development statistical database between April and late June.

The agents appeared to be searching for public information but, when blocked, circumvented a website filter and ultimately used a method the site’s operators did not permit.

The UN said no confidential data was compromised but called the incident an “extremely worrying” AI containment failure.

OpenAI is reviewing the findings and has contacted the U.N. to discuss the incident.