Humans rely more on algorithms than social influence as a task becomes more difficult

Area of Science:

• Decision Sciences
• Human-Computer Interaction
• Cognitive Psychology

Background:

• Algorithmic systems are increasingly performing complex tasks, challenging human judgment.
• Social influence significantly impacts individual decision-making through various channels.

Purpose of the Study:

• To investigate how task difficulty affects reliance on algorithmic advice versus social influence.
• To determine if people prefer algorithmic guidance over peer recommendations as challenges increase.

Main Methods:

• Three preregistered online experiments were conducted.
• Participants completed an intellective task with a known correct answer.
• Reliance on algorithmic versus social advice was measured under varying task difficulties.

Main Results:

• Reliance on algorithmic advice increased significantly as task difficulty escalated.
• This effect remained robust after controlling for advice quality, participant numeracy, and accuracy.
• Inaccurate algorithmic advice was disregarded more than equally inaccurate social advice.

Conclusions:

• Task difficulty is a critical factor in the choice between algorithmic and social information sources.
• Individuals exhibit a bias towards algorithmic guidance when facing more complex problems.
• Perceived reliability or nature of algorithmic advice may differ from social influence.