News for hackers. Stuff that matters.
Re:Kenyan moderators decry toll of training of AI by dstroot on Thursday August 03, 2023 @12:05PM (#36987742)
What is stopping us from creating a "supervisory AI" that understands the difference between good content and bad content and having it perform the role of these humans? More broadly speaking training and overseeing a sufficiently advanced AI seems like it could only be done effectively by another AI. The human oversight could be temporarily redirected to validating the supervisory AI.

There are no comments on this comment yet.