[cctalk] Re: AI applied to vintage interests

On 1/18/23 04:16, Peter Coghlan via cctalk wrote: technology, it's hard to imagine them managing to take over the world > before they crash/panic/BSOD/bomb etc depending on their specific > flavour and/or get bricked by malware and/or fail due to > leaking/dried up electrolytic capacitors and/or batteries. It's hard > enough to keep systems going without interruption when this is what > people are actively trying to do. > > In any case, how are they going to prevent the humans from cutting > off their power? > Just imagine, instead of a soldier or airman sitting at a display picking out targets for a airborne drone, doing away with the psychological stress and uncertainty and letting an AI select and attack targets.   As far as I can tell, this doesn't violate any conventions.  How about artillery using the same system?   The AI will be distant from the actual weapon, so no concerns about cutting off power. Of course, I'm stating the obvious--I would be very surprised if various governments weren't already developing platforms based on AI. --Chuck