OpenAI CEO to Push for Global AI Safety Standards at the UN

Beating AI News Flash: OpenAI CEO Sam Altman will speak at the UN Security Council, advocating for unified global AI safety standards while avoiding restrictive measures that could slow innovation. OpenAI stated that it does not support the establishment of a specific international regulatory body, preferring basic safety standards plus democratic oversight. Anthropic CEO Dario Amodei is also expected to participate via video.

Previously, according to FT, OpenAI on Monday called for the US government to lead the establishment of international AI standards and suggested that countries coordinate to assess the capabilities and risks of AI models. OpenAI CEO Sam Altman will present the relevant proposal to the UN Security Council this week. OpenAI specifically warned that recursive self-improvement (RSI) may bring new risks, namely that more and more AI model training work is being done by other AI systems. Without appropriate safety measures, AI could rapidly iterate to a degree beyond which humans can effectively supervise, causing humanity to gradually lose actual control over AI development. The relevant proposal comes as concerns about AI safety are rising.