Public attitudes on performance for algorithmic and human decision-makers
2024

Public Preferences for Algorithmic vs. Human Decision-Makers

Sample size: 9000 publication 10 minutes Evidence: high

Author Information

Author(s): Bansak Kirk, Paulson Elisabeth

Primary Institution: University of California, Berkeley

Hypothesis

How do public preferences for algorithmic and human decision-makers differ based on performance metrics?

Conclusion

The study found that while respondents generally prefer human decision-makers, they prioritize efficiency over fairness in their evaluations.

Supporting Evidence

  • Efficiency was the most important performance metric in evaluations.
  • Respondents showed a general preference for human decision-makers over algorithms.
  • Fairness was the least prioritized metric among respondents.
  • Preferences varied significantly across different demographic groups.

Takeaway

People like humans more than robots making decisions, but they care more about how well those decisions work than about fairness.

Methodology

The study used a conjoint experimental design where respondents evaluated pairs of decision-maker profiles in high-stakes scenarios.

Potential Biases

Potential biases in public perception of algorithmic decision-making were not fully explored.

Limitations

The study only considered two scenarios and one definition of fairness.

Participant Demographics

The sample was representative of the US population in terms of gender, age, region, race/ethnicity, household income, and education.

Statistical Information

P-Value

p<0.0001

Statistical Significance

p<0.0001

Digital Object Identifier (DOI)

10.1093/pnasnexus/pgae520

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication