AI advice suppresses people's willingness to say "I don't know", even when the advice is wrong and accuracy is incentivized

Five experiments involving 3,132 participants demonstrate that AI advice suppresses the human willingness to say "I don't know", even when accuracy is incentivized. By providing fluent but incorrect answers, AI assistants nearly eliminated participants' tendency to suspend judgment, regardless of whether advice was requested or displayed. Consequently, participants answered more questions but were correct only about a third as often as those without AI access, despite maintaining high levels of confidence in their responses.