Saw this being discussed on Blossom earlier, recent reporting suggests the U.S. military used Anthropic’s Claude AI model in connection with a Venezuela-related operation.
Even if the AI’s role was limited to analysis or intelligence support, it marks a notable shift: commercially developed large language models being integrated into national security.
As generative AI tools become more capable, their use in military and intelligence contexts may expand.
https://www.wsj.com/politics/national-security/pentagon-used-anthropics-claude-in-maduro-venezuela-raid-583aff17
5 Comments
This seems like a significant milestone in the evolution of AI deployment. If commercial AI models are increasingly used for intelligence analysis, logistics planning, or operational support, we may be entering a phase where defence infrastructure becomes tightly coupled with privately built AI systems.
Historically, major technological shifts in warfare, from satellites to cyber capabilities, have reshaped geopolitical strategy. Could widespread adoption of advanced AI tools create a similar structural change in how military power is exercised over the next decade?
»Hey Gemini, should we bomb Iran?«
Waagh-Minister Pete Kegbreath
We don’t have any detail on how it was used, so anything said about this is no more than baseless speculation.
There are obvious potential safety concerns, but again: WE. DON’T. KNOW. ENOUGH. There are tons of dangerous ways of using AI, but there also are tons of perfectly safe and small ways of using it.
This is no more than a personal rant: Of course, some people will LOVE to strategically ignore the “boring” options and build a castle of ideas and propositions based on this smoke of a declaration of Claude being used in this. I despise that kind of speech and narrative.
“reportedly used” means hearsay, we have no idea to what extent that’s true, if true at all.
If true, we should definitely congratulate the military for the fast adoption of the cutting edge technologies. I think the future of AI in warfare is clear – we’ll have more of it at every stage, from data analysis to planning and execution. The defense infrastructure has always been tightly coupled with privately built systems.
How military power is exercised? I think autonomous weapons may lower the barrier to engagement. The most important consideration when making a decision to engage is the estimated number of casualties on our side, and if that’s zero ..
Revisit this question when the first massive terrorist attack is compiled and executed using commercial AI software.