It might be my inner Luddite talking, but LLM use in defense and intelligence terrifies me. What happens when built-in model biases or hallucinations affect hum

It might be my inner Luddite talking, but LLM use in defense and intelligence terrifies me. What happens when built-in model biases or hallucinations affect human safety? Who is to blame and how will this be mitigated? Fascinating but scary.