An experimental study suggests that people are less likely to behave in a trusting and cooperative manner when interacting with AI than when interacting with other humans. Scientists use experimental games to probe how humans make social decisions requiring both rational and moral thinking.