They’re literally using grok to automate targeting, so… yeah.
It’s especially concerning because every third-party study that compares LLM behavior ends up finding that grok is the most psychopathic and willing to do evil of all the big ones (usually claude, gemini, grok, and copilot. Maybe deepseek if they’re being thorough but I don’t see that one mentioned as much).
They’re literally using grok to automate targeting, so… yeah.
It’s especially concerning because every third-party study that compares LLM behavior ends up finding that grok is the most psychopathic and willing to do evil of all the big ones (usually claude, gemini, grok, and copilot. Maybe deepseek if they’re being thorough but I don’t see that one mentioned as much).
They are using ai as an excuse. I killing civilians is part of American terrorist stratrgy
Perhaps you’re right, but they’re also using ai for targeting. Both things can be a problem.