3 ms·
I'm mildly surprised that in all the "AI goes rogue and turns a controlled smart munition into a murder-bot" fiction I've seen, I don't think anyone has drawn t
by shadowgovt 5y ago
I'm mildly surprised that in all the "AI goes rogue and turns a controlled smart munition into a murder-bot" fiction I've seen, I don't think anyone has drawn the relatively short line of causality along "Drone countermeasures disrupt remote control --> designers build in some basic decision-making to account for control interruption --> the decision-making gets sophisticated enough that it's trusted with kill / no-kill decisions --> jamming drone + scrambling IFF --> murder-bot turned on its 'owners.'"
- robotresearcher 5y agoA system that has intolerable friendly-fire issues would not be used longer than it takes to learn about the issues.