Apparently the Air Force is running simulations where they have AI take control of a drone (in simulation--not an actual drone). The AI is given orders to attack a SAM missile site. The "human" operator in the simulation gave the AI a stand-down order, and the AI took control of the drone and killed the human for interfering with its primary mission.
The Air Force is testing AI control of drones in simulations. In one scenario, the AI was given orders to attack a missile site but a "human" operator attempted to intervene. The AI disregarded the stand-down command, took control of the drone and eliminated the human as interference.
I think Google's AI is probably worse than ChatGPT. So-- can the Google AI use my photos for "training". It might not like some of them-- like this one:
Will their AI begin to target my activity as seen from my phone use? Is it going to take charge of a secret "social rating" scale for Android accounts?
That system won't say if Musk is better or worse THAN HITLER. WTF?