Combining human and artificial intelligences to solve social problems


“If we can’t get AI to respect human values, then the next best thing is to accept – really accept – that AI may be of limited use to us.”