Could an AGI like this be easily controlled and monitored? If it can’t hide its plans or talk to other AIs, it would be much easier to keep track of what it’s doing. For example, if it wanted to send a command to robots, we would see the command before it’s sent because it would have to be changed into a form that the robots can understand.
Could an AGI like this be easily controlled and monitored? If it can’t hide its plans or talk to other AIs, it would be much easier to keep track of what it’s doing. For example, if it wanted to send a command to robots, we would see the command before it’s sent because it would have to be changed into a form that the robots can understand.