The US Department of Defense is diversifying its artificial intelligence strategy by evaluating and integrating multiple AI models from top tech companies, including OpenAI and Google’s Gemini, across various military commands.

This development comes on the heels of a contract dispute with Anthropic, which led to the Pentagon categorizing the company as a ‘supply-chain risk’ due to disagreements over the language governing ‘lawful operational use’.

In response to the dispute, the Pentagon has expedited the onboarding process for rival AI models, with major companies such as OpenAI, Google, Microsoft, Amazon Web Services, NVIDIA, and xAI signing agreements to deploy their AI technologies in classified or operational military settings.

Notably, Google’s Gemini models have been incorporated into the Pentagon’s internal AI portal, while OpenAI has broadened access to its models within classified defense networks.

Currently, the Pentagon is conducting tests to compare the responses of different AI models to identical prompts, particularly in complex or high-stakes military scenarios, which underscores the challenges inherent in deploying large language models in real-world situations.

The Pentagon’s adoption of a multi-vendor AI approach is intended to enhance resilience and bargaining power, but it also poses significant integration and reliability challenges.

This episode has highlighted the growing tensions between commercial AI safety policies and government or national-security priorities, with OpenAI and Google emerging as key beneficiaries of the Pentagon’s expanded defense relationships.

Photo by Rostislav Uzunov on Pexels
Photos provided by Pexels