On the same day Anthropic filed its lawsuit against the Department of Defense, Bloomberg reported that Google is deploying Gemini AI agents across the Pentagon's three million-person workforce to automate routine unclassified tasks.
Google's deployment — under the direction of Emil Michael, the under secretary of defense for research and engineering — will initially operate on unclassified networks, handling administrative and workflow automation that doesn't touch classified data or weapons systems. That's an important technical qualifier, but it doesn't resolve the fundamental tension the day's news surfaces: two of the most powerful AI companies in the world are now on opposite sides of a conflict over what the American military is allowed to do with AI.
Anthropic's position — that some use cases are simply off the table regardless of who's asking — represents a more restrictive view of what responsible AI deployment looks like. Whether that position is principled, commercially sustainable, or both, is what this week is beginning to test.
Comments