Scientists find AI's social desirability bias exceeds human standards. What does this mean for AI ethics and human interactions?
