News US Military Drone AI Simulation Reportedly Turned on Its Human Operator

[MarineRecon]

Yes, this did not happen.
https://skeptics.stackexchange.com/...he-human-operator-in-a-simulation-environment

Uh,oh!

 

[bit_user]

If you have any semblence of an historically accurate understanding of political history from the last 90 years, then you know where all this can, and may be, heading.
And its not good.

Not exactly sure what you're getting at. Perhaps it's something like this?

Consider that the ability of bad actors (i.e. rogue states, terrorists, & organized crime) to do bad things has been practically limited by how many competent, amoral underlings they can attract to their organization and keep in line. A real concern I have is that with AI bots capable of performing complex tasks, suddenly you now have something like the "robot army" scenario of a comic book supervillain.

 

[velocityg4]

While I know this was just a thought experiment. It seems you give the system equal points to following a command as to making a kill plus negative points to disobeying. It would follow the command as the command is guaranteed points while a kill isn't guaranteed (it could miss).