Anthropic snubs Pentagon push to open Claude AI for unmanned weapons systems or mass surveillance use

[Redaktion]

Anthropic is likely to remain the leading AI company when it comes to the safety and social responsibility of its models. Despite all the potential repercussions, it has turned down the Pentagon's request to remove the guardrails against military use and surveillance inherent to its Claude AI agent.

https://www.notebookcheck.net/Anthropic-snubs-Pentagon-push-to-open-Claude-AI-for-unmanned-weapons-systems-or-mass-surveillance-use.1236524.0.html

[Admiral Snackbar]

They defend "democracy" with the heavy handedness of a totalitarian dictatorship.

[jdrch]

I read both the Pentagon and Anthropic statements. Basically, the Pentagon wants AI for all lawful use, while Anthropic wants specific guardrails against domestic surveillance and pure AI killchains.

This is fascinating as, AFAIK, it's the 1st instance of a major private company standing up against this administration. Good on them.

[Admiral Snackbar]

This is fascinating as, AFAIK, it's the 1st instance of a major private company standing up against this administration. Good on them.

The pentagon wants AI to help it kill people and Anthropic wants to market their product to earth conscious business hippies. This clash is a great marketing campaign for Anthropic. You could pay for this kinda publicity.

And really, can't the pentagon just use OpenAI, Google, Meta, or xAI to help it kill people? Any one of them would jump at the opportunity.