OpenAI gives US government early access to new frontier AI models

OpenAI is providing the US AI Safety Institute at the National Institute of Standards and Technology (NIST) early access to its next frontier AI model. The collaboration aims to improve AI evaluation methods, says OpenAI CEO Sam Altman. He also reaffirmed the company's pledge to dedicate at least 20% of its computing resources to safety measures. This comes after former OpenAI safety researcher Jan Leike claimed the company had not kept this commitment. Altman also said that in May, OpenAI removed clauses that allowed the company to revoke previously granted stock options and required employees to remain silent. The change aims to create an environment where employees can "raise concerns and feel comfortable doing so," Altman writes.