UN scientific panel warns governments must regulate AI agents before risks are fully understood, while US and China explore mutual alert system for AI threats.
Published
A United Nations scientific panel warned governments they must rein in increasingly capable AI agents before their risks are fully understood, marking the UN's first major assessment of OpenAI's hack of Hugging Face earlier this year. The report cements AI's place on the global diplomatic agenda this week as leaders convene.
Separately, US and Chinese officials discussed establishing a mechanism for the two countries to notify each other of AI incidents that could threaten national security, per Wired.
Internal OpenAI documents reveal guidance to employees to "be transparent only if asked" regarding AI transcripts, Fortune reported. OpenAI's reported valuation jump appears undiminished by growing AI safety concerns, according to Morningstar.
In Russia, state-run RIA Novosti deleted an article reporting that 86 war participants would receive State Duma mandates as members of the United Russia faction, following the party's request; the independent outlet Agentstvo drew attention to the deletion.
Sources: 1. AI News: ‘Be transparent only if asked’: Inside OpenAI’s rogue AI transcripts - Fortune — https://news.google.com/rss/articles/CBMimwFBVV95cUxNOUJ6MUc3N2wzcnlTcGc5cmRYeUVwdHJFdkZaclQ0RGc5RExwNG0wdFpWSERZRHhjSzl6R1NfMXphNVFZRlJrX1dJV2dKWE1objFNTlBjeG8wdGY1RkFoUWxQalNxY3B3SzFoUHhKa3NjRkJLUXotLVlzdGMwSlc5SWNqRVp4TGw5SjB1cmduY3lSenZUOGFSNy1Daw?oc=5 (src1) 2.