五角大楼希望AI公司扩大对机密军事网络的接触,但公司抵制安全顾虑。 The Pentagon wants AI firms to expand access to classified military networks, but companies resist over security concerns.

五角大楼正在推动大型AI公司,如OpenAI和人类学公司,扩大在机密军事网络上使用其工具的机会,目的是将先进的AI纳入敏感的国防行动。 The Pentagon is pushing major AI firms like OpenAI and Anthropic to expand access to their tools on classified military networks, aiming to integrate advanced AI into sensitive defense operations. 虽然目前的准入仅限于非机密系统,如300万人员使用的genai.mil平台,但军方寻求更广泛的部署,用于特派团规划和确定目标。 While current access is limited to unclassified systems, such as the genai.mil platform used by 3 million personnel, the military seeks broader deployment for mission planning and targeting. 科技公司仍持谨慎态度,以AI错误和滥用等风险为借口,而人类则抵制某些军事用途。 Tech companies remain cautious, citing risks like AI errors and misuse, with Anthropic resisting certain military uses. OpenAI维持其目前的协议仅限于非保密用途。 OpenAI maintains its current agreement is restricted to unclassified use. 这一努力突显了对迅速采用AI的军事要求与技术公司对减轻风险的关切之间日益紧张的关系。 The effort highlights growing tension between military demands for rapid AI adoption and tech firms’ risk mitigation concerns.