Google Grants Pentagon Access to AI Models

Google has expanded the Pentagon’s access to its artificial intelligence capabilities, granting API access to its commercial AI models for classified work. This move follows Anthropic’s refusal to grant the Department of Defense similar access with restrictions, leading the DoD to designate Anthropic as a “supply-chain risk” and initiate a lawsuit.

The agreement, finalized on Monday, allows the Pentagon to directly integrate Google’s AI software into classified military environments without requiring custom-built models, according to a person familiar with the matter. This approach aims to support national security needs while aligning with existing industry practices. The deal was reported by GuruFocus and further detailed by TechCrunch.

Internal Concerns and Previous Disputes

The expansion of access hasn’t been without internal dissent at Google. Hundreds of AI researchers reportedly sent a letter to CEO Sundar Pichai urging the company to avoid providing its systems for classified defense workloads. These researchers expressed concerns that such technologies could centralize power and introduce risks through potential errors.

This situation echoes past friction within Google regarding its involvement with defense projects. In 2018, employee protests led the company to discontinue its participation in Project Maven, a Pentagon contract focused on drone footage analysis, despite Google’s initial assertion that the work was intended for non-offensive purposes.

Anthropic’s Stance and the DoD Response

Anthropic’s decision to resist the Pentagon’s demands for unrestricted AI access triggered the current situation. The company sought to implement safeguards to prevent its AI from being used for domestic mass surveillance and the development of autonomous weapons. The DoD, however, desired unrestricted use of the technology.

As a result of Anthropic’s refusal, the Department of Defense labeled the company a “supply-chain risk” – a designation typically reserved for foreign adversaries. Anthropic subsequently filed a lawsuit against the DoD, and last month, a judge granted the company an injunction against the designation while the case proceeds.

Google Follows OpenAI and xAI

Google is now the third AI company to pursue a deal with the DoD following Anthropic’s stance. OpenAI and xAI both signed agreements with the Department of Defense shortly after Anthropic’s refusal to grant unrestricted access.

Google’s agreement includes language stating the company does not intend for its AI to be used for domestic mass surveillance or in autonomous weapons, similar to the contract language with OpenAI. However, The Wall Street Journal reports that the legal enforceability of these provisions remains unclear.

Employee Opposition and Lack of Comment

Despite the agreement, significant opposition remains within Google. An open letter signed by 950 employees urged the company to follow Anthropic’s lead and refrain from selling AI to the Defense Department without similar safeguards.

Google has not yet responded to requests for comment regarding the deal or the employee concerns. The company’s decision highlights the growing tension between the potential benefits of AI for national security and the ethical considerations surrounding its use in military applications.

The timing of this agreement also intersects with a broader reassessment within the Pentagon’s AI ecosystem, as the DoD explores alternative partners following the breakdown in negotiations with Anthropic.