When AI Entered the Battlefield And Why the Pentagon Pushed Claude Out
AI is becoming central to modern warfare. Data analysis platforms and defence tech companies will likely see long-term growth globally.
For decades, the biggest challenge for modern militaries wasn’t weapons — it was information.
After 9/11, the United States began collecting massive intelligence data from drones, satellites, and surveillance systems. The problem? Humans simply couldn’t process it all.
At one point, a single drone required 19 analysts just to watch the live footage.
So the Pentagon turned to technology.
First came data-analysis platforms like Palantir, which could connect thousands of databases and detect patterns in intelligence data. These systems helped identify threats such as insurgents converting garage door remotes into bomb detonators.
But these systems had a limitation — they could find patterns, not explain them.
To solve this, the US military launched Project Maven, an AI initiative designed to analyse battlefield data using machine learning.
By 2024, the next step arrived: Large Language Models (LLMs).
Anthropic’s AI model Claude was integrated into Pentagon systems through Palantir’s platform. The idea was simple:
Instead of analysts navigating complex systems, they could ask questions in plain English and receive intelligence reports instantly.
However, conflict soon emerged.
Anthropic maintained strict rules about how its AI could be used — especially around weapons, cyber operations, and mass surveillance.
• Bharat Electronics (BEL)