algoriAlgorithms are playing a bigger role in decisions that shape people’s lives, from pretrial release to bank loans. Critics argue that relying on algorithms locks in bias and sacrifices fairness. But when faced with a choice, people seem to value efficiency over fairness.

In a study from UC Berkeley, 9,000 Americans were asked to choose between decision-makers for two tough calls: whether to grant bail and whether to approve a loan. Options included two humans, two algorithms, or one of each, with simulated performance stats provided for each choice.

Efficiency wins

Across groups—whether divided by race, education, political leaning, or views on artificial intelligence—participants picked the option that was most efficient, no matter if it was a human or an algorithm. That said, people leaned slightly toward human decision-makers, with Republicans showing a stronger preference for humans than Democrats.

This focus on efficiency came despite many participants claiming that fairness was one of their top priorities. The researchers noted that people’s actions didn’t match their words. They suggest that as algorithms prove themselves more effective, resistance to using them will likely fade, no matter the political or cultural background.