Google、Microsoft、xAIが米政府の審査にモデルを提出
Google DeepMind、Microsoft、xAIは、商務省のAI標準・イノベーションセンター(CAISI)がリリース前に新モデルを審査することに同意した。同センターは2024年以降、主にOpenAIやAnthropicとともに40件の審査を実施しており、今後は新たなパートナーも含めて拡大する。
Google DeepMind, Microsoft, and xAI will subject new frontier AI models to pre-deployment review by the US Commerce Department's Center for AI Standards and Innovation (CAISI). CAISI, which has conducted 40 reviews of OpenAI and Anthropic models since 2024, aims to assess capabilities and national security implications prior to public release. This expansion aligns with President Trump’s AI Action Plan, as the administration considers an executive order to formalize oversight of advanced AI systems.