Daniel Kokotajlo comments on Do any of the AI Risk evaluations focus on humans as the risk?