My thoughts were on this line also... the same reason the crowd was shocked when Lee Sedol made an unexpected move in game 4. [0]

Is there any research done to prove/disprove that groupthink is inherently worse compared to a single genius? (less risky behavior, so less risk/less reward??) Are we training AIs that will be safe, not bold independent thinkers?

0: https://www.wired.com/2016/03/two-moves-alphago-lee-sedol-re...