Sam Altman Backs Proposal to Temporarily Halt Advanced AI Development

OpenAI CEO Sam Altman has expressed support for the proposal to slow down the development of the most advanced artificial intelligence systems for a period of time. He stated that, if necessary for humanity’s safety, OpenAI is prepared to shut down all its GPUs. However, he noted that the likelihood of such a scenario happening is only a reasonable estimate.

Altman endorsed the proposal made by Anthropic CEO Dario Amodei, which called for a temporary reduction in the pace of advanced AI development to ensure security and responsibility. He shared these views during an interview with Fortune.

Regarding independent safety testing, Altman revealed that OpenAI will have its AI systems evaluated by independent organizations. He committed to providing these testers with the same level of access and information as OpenAI employees. The independent evaluators will assess whether OpenAI’s AI systems pose risks to humans and whether the company is making sufficient efforts to mitigate those risks.

The phrase “shutting down GPUs” refers to turning off the graphics processing units required to build and operate AI systems. GPUs play a crucial role in processing large amounts of data. Therefore, when Altman says OpenAI is “ready to shut down all GPUs,” it means that if risks become excessive, OpenAI is prepared to pause rapid AI development and deactivate its computing infrastructure. He further emphasized that to ensure AI safety and protect humanity’s interests, it may be necessary at times to slow development repeatedly through such agreements, not just occasionally.