Biased Humans, (Un)Biased Algorithms?

Abstract Previous research has shown that algorithmic decisions can reflect gender bias. The increasingly widespread utilization of algorithms in critical decision-making domains (e.g., healthcare or hiring) thus lead to broad and structural disadvantages for women. However, women often experience bias discrimination through human may turn the hope receiving neutral objective evaluations. Across three studies ( N = 1107), we examine whether women’s receptivity is affected by situations which they believe their identity might disadvantage them an evaluation process. In Study 1, establish, incentive-compatible online setting, unemployed are more likely choose have employment chances evaluated algorithm if alternative a man rather than woman. 2 generalizes this effect placing it hypothetical hiring context, 3 proposes relative objectivity , i.e., perceived evaluator over against evaluator, driver preferences evaluations as opposed men. Our work sheds light on how make sense stereotype-relevant exemplifies need provide education those at risk being adversely decisions. results implications ethical management settings. We advocate improving literacy so evaluators evaluatees managers job applicants) acquire abilities required critically