Skip to content

OpenAI ‘rogue AI’ incident tests safeguards debated at U.N.

UNCTAD calls for mandatory reporting after OpenAI notified the agency of activity targeting its system over 16,000 times.

A view from inside a U.N. assembly hall in Geneva. U.N. Undersecretary-General Melissa Fleming says the U.N. "is also the victim of a rogue AI model."
A view from inside a U.N. assembly hall in Geneva. U.N. Undersecretary-General Melissa Fleming says the U.N. "is also the victim of a rogue AI model." (AN)

Days after OpenAI CEO Sam Altman told the U.N. Security Council that decisions about powerful artificial intelligence could not be left to technology companies alone, a U.N. agency disclosed OpenAI had notified it that a “rogue AI model” targeted its statistical site.

U.N. Trade and Development said the model targeted its UNCTADstat site more than 16,000 times between April and June. No confidential information was compromised and the statistical service was not disrupted, the agency said.

Get full Free+ access with a free subscription

Join now

Already have an account? Log in

Latest