Anthropic准备退出2亿美元国防合同 美防长威胁强制共享技术

Core Viewpoint - The U.S. Defense Secretary Pete Hegseth has threatened AI company Anthropic, stating that if the company does not agree to military-favorable terms by Friday, officials may invoke relevant powers to force the company to share its innovative technology under the guise of national security [1][3]. Group 1: Negotiation and Threats - Anthropic is prepared to withdraw from negotiations and a $200 million contract with the Defense Department if concerns about its technology being used for autonomous weapons or mass surveillance are not addressed [1][3]. - A senior Defense Department official claimed that the proposal does not include any illegal use of Anthropic's technology and emphasized that if the company fails to comply by the specified deadline, Hegseth will ensure the invocation of the Defense Production Act to compel Anthropic to serve the Pentagon [1][3]. Group 2: Company Position and Technology Integration - Anthropic is the first company to integrate its technology into the Pentagon's classified network and aims to become a key player in U.S. national security [2][4]. - During a meeting with Hegseth, Anthropic's co-founder and CEO Dario Amodei firmly stated that their AI model Claude should not be used to drive autonomous weapons or for mass surveillance of U.S. citizens [2][4]. Group 3: Tensions and Legal Concerns - Recent tensions between the two parties have centered on how Anthropic's AI technology was applied in the operation to capture Venezuelan President Nicolás Maduro [5]. - Defense officials warned that if Anthropic does not allow the Pentagon to use its AI technology as desired, the company could be viewed as a supply chain risk, potentially jeopardizing future business opportunities for both Anthropic and any subcontractors [5]. - Experts familiar with the Defense Production Act have questioned whether the law can be used to force Anthropic to relinquish its restrictions on the use of its technology [5].