Odaily Planet Daily reports that the U.S. government is engaged in high-level negotiations with AI companies, planning to announce voluntary standards for releasing new models as early as next week.
This standard will establish benchmarks for models with advanced network capabilities and set a release schedule to streamline future model deployments. This move aims to implement the AI executive order announced last month by the U.S. President, and U.S. officials will also clarify access to advanced AI models both domestically and internationally, potentially laying the groundwork for a global framework that includes U.S. allies.
Previously, the U.S. government intervened in the release of new tools by Anthropic and OpenAI; Anthropic’s latest model was subject to export controls on June 12 due to concerns such as cybersecurity risks, and these restrictions were lifted on Tuesday. Currently, AI lab technical teams regularly meet with government officials to reach consensus. (FT)
