Computer science has long understood what it takes to keep AI under control

Recent AI hacking incidents revive a long-standing computer science problem: what happens when machines pursue a goal without enough limits.

Opens the original publisher in a new tab. Full articles belong to their publishers; a subscription may be required.