Governments, especially the FED, are racing to bring new AI tools into their military systems. At the same time, the companies building this stuff are showing some reluctance at the desired pace of the uptake...
Take Anthropic, for example - they just signed a $200 million deal. But they’re laying down some ground rules: no mass surveillance, and definitely no fully autonomous weapons systems operating without a human watching.
They are trying to layer ethics into the deployment, but the FED isn't that keen....
The problem is, despite Anthropics efforts, tech companies don’t get the final say on how their stuff is used - governments do.
Once the tech changes hands or gets integrated into military systems, it stops being about corporate ethics and starts being about military doctrine. And even if a company tries to bake rules into their contracts, good luck policing them, especially when the details are classified.
Companies like Anthropic still have some bargaining power, at least for now. Their AI models are advanced and not easy to copy, and it's hard for the Pentagon to just swap them out without losing something important.
But if another firm comes along with fewer scruples, governments might ditch the cautious players and go with someone less picky. In a world where everyone wants the best tools, the company insisting on ethical constraints starts looking like an outlier, maybe even a liability.
Maybe Anthropic’s stand seems resolute now. But the big question is whether that approach can survive when competition heats up—and when there’s always someone out there willing to do what others won’t.
