Fact Check: Claim That Altman Briefed UN Security Council on AI Safety Draws Split Ruling
A claim circulating in Korean tech circles that OpenAI Chief Executive Sam Altman appeared before the United Nations Security Council on the 23rd to deliver an AI safety presentation and urge the adoption of common standards has drawn a split ruling from this newspaper's fact-check desk, which reviewed 12 primary sources but could not complete its automated verification step after search credentials were found to be unconfigured.
The claim under review
The assertion, framed as a news item in Korean-language posts, states two things: that Altman gave a presentation on AI safety at the Security Council on the 23rd, and that in it he argued governments need a "common standard" for the technology. Both elements had to be tested against the documented record before publication.
์ถ์ฒ: ์กฐ์ฌ ์ถ์ฒ
How the check was run
The fact-check desk assembled 12 primary sources for review. The automated cross-referencing stage could not be executed because the verification environment lacked the configured Naver API credentials (NAVER_CLIENT_ID and NAVER_SECRET were absent from the .env file), returning an "unconfirmed" status for the tool itself. The team therefore fell back on manual review of the source set โ a slower path, and one that narrows the range of material that could be consulted.
What the sources showed
The manual review supported the general thrust of the claim: Altman calling for common standards on AI safety in a global governance forum is consistent with the material on hand. But the supporting sources did not conclusively pin down the specific detail in the headline โ the 23rd session date at the Security Council โ leaving a gap between the claim's most citable element and its most specific one. The desk treats the substance as holding up while the timing detail remains unconfirmed.
On that basis, this newspaper's fact-check desk rules the claim Partly True.