Les États américains appellent à la participation aux tests de modèles d'IA, renforçant la coopération en cybersécurité

Gold Financial reports that on May 6th, senior security officials from more than ten states in the United States urged artificial intelligence companies to include these states in frontier model testing, citing that developers are currently more inclined to collaborate with Washington and large tech companies.
In a letter sent on Tuesday to the CEOs of OpenAI, Anthropic, Microsoft, and Google, a group of state chief information security officers, homeland security advisors, and technical officials stated that the states are still excluded from early access programs. The letter read: “While we appreciate the launch of new cyber defense projects, we are deeply concerned about the current practice of prioritizing collaboration with federal partners and large tech companies, while essentially excluding cooperation with state and local agencies responsible for providing and protecting many of the nation’s critical services.” (Jinshi)