@blogdiva Just the concept of an AI directing bombs puts me in mind of the #StarTrek episode "A Taste of Armageddon," where the AIs on two sides simulated a war, but the casualties still had to be real. So for each simulated strike, the corresponding number of people were executed.
Nearly 60 years after that episode, and we're almost there.
@blogdiva no wonder they fired all the ML ethics teams. Those teams wouldn’t even sign off on LLMs, no way in hell they’d…
I’m trying to find the link to an article I remember a few years back that talked about some sort of leaked presentation from the DOD or CIA that a contractor? maybe gave about facial recognition software to track potential “enemy combatants”. The relevant agency said they didn’t use it as the error rate was too high and that it was all speculative, etc.
I’m really, really curious about the specific details of those slides now.
Add comment