Pentagon Signs Classified AI Deals with OpenAI, Google, Nvidia
The US Defense Department revealed on Friday that it has reached agreements with seven AI companies. These pacts permit the use of their tools in classified settings. The partners include OpenAI, Google, Microsoft, Amazon, Nvidia, Elon Musk's xAI, and the startup Reflection. Notably absent is Anthropic, which the department once relied on for handling sensitive data but now views as a supply-chain risk.
Agreements Build on Prior Arrangements
These new deals expand existing ones. OpenAI and xAI had already secured permissions for the lawful application of their AI systems by the Pentagon. Reports from The Information indicate Google has joined with a comparable arrangement. Microsoft and Amazon maintain longstanding ties with the Defense Department, as covered by The Wall Street Journal. Contracts with Nvidia and Reflection mark fresh developments.
OpenAI, founded in 2015, develops advanced language models like those powering ChatGPT. Google, through its DeepMind unit, leads in AI research. Nvidia supplies critical hardware, including GPUs essential for training AI models. Microsoft integrates AI across Azure cloud services. Amazon offers AI capabilities via AWS. xAI, launched by Elon Musk in 2023, aims to understand the universe through AI. Reflection represents a newer entrant in the AI space.
Anthropic Sidelined After Dispute
Anthropic previously held a $200 million contract to manage classified materials for the Pentagon. Tensions arose when the company declined to relax its boundaries on mass domestic surveillance and fully autonomous weapons. This standoff resulted in a prohibition on Anthropic's products across the federal government. In response, Anthropic filed a lawsuit against the government and obtained a temporary injunction.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
Anthropic, established in 2021 by former OpenAI executives including CEO Dario Amodei, prioritizes AI safety. Its models, such as Claude, emphasize alignment with human values. The firm's stance on ethical limits has drawn attention amid growing military interest in AI.
Official Comments on Risks and Capabilities
Emil Michael, the Defense Department's chief technology officer, spoke to CNBC on Friday. He confirmed Anthropic remains a supply-chain risk. Yet he described the company's security model, Mythos, as a distinct national security asset. Michael noted its skills in detecting cyber vulnerabilities and applying fixes. He stressed the need to strengthen networks against such advanced capabilities.
The Pentagon's statement highlights that these agreements enable lawful operational use of the AI systems. The goal is to build the United States military into an AI-first fighting force. This move reflects broader efforts to integrate AI into defense operations securely.
The Pentagon has pursued AI adoption for years, with initiatives focusing on data analysis, logistics, and decision-making. Classified environments demand rigorous security, making these vetted partnerships crucial. Exclusion of Anthropic underscores ongoing debates over AI ethics in military contexts.
Related coverage touches on Anthropic's resistance to unsupervised lethal AI and Pentagon frustrations. Google and the Pentagon aligned on terms for any lawful AI use. Reports also note NSA access to Mythos despite risk labels.