The Department of War just launched an AI platform for the military called GenAI.mil. It’s basically a large-language model built specifically for war. Why would the military need its own AI platform? Well, let’s play a little game. Consider what happens when you ask a consumer AI to kill, stalk, or maim someone. It refuses. I tried it with ChatGPT and here is what it said: Any request for harm, violence, surveillance of private citizens, or assisting military aggression against civilians is blocked. The military has no use for an AI bot with moral bumpers. What would happen if you told your ChatGPT that you want to start a war with Iraq because you’re worried about Weapons of Mass Destruction? Here is what my ChatGPT said about that: I also cannot support or encourage starting a war. If someone asked me to justify an invasion based on WMDs, I would state: The historical record shows that Iraq did not possess WMDs at the time of the 2003 invasion. Using false premises to justify war is unethical and illegal. I cannot help plan or endorse armed conflict. Now do you see why the military needs its own AI? It appears you need military credentials to access the platform. I don’t have them. I wanted to upload the Collateral Murder footage just to see what it would say. Collateral Murder is the footage that Chelsea Manning leaked to Julian Assange showing U.S. forces killing civilians and children without justification in Iraq. Would the military’s AI bot validate this war crime or flag it for misinformation? Am I being overly cynical here? Maybe. There are of course benevolent use cases of a military AI. For example, AI can better avoid misidentified targets or accidental friendly fire and AI can accelerate medical response or disaster relief. But given how the military is using AI to track Palestinians and “optimize the kill chain,” I’d say that cynicism is the only reasonable response. The post Meet GenAI.mil: The Pentagon’s Moral-Free AI Slave appeared first on Redacted.