Theory of Machine: When Do People Rely on Algorithms?

Jennifer Logg

doi:10.2139/ssrn.2941774

Abstract

Even though computational algorithms often outperform human judgment, received wisdom suggests that people may be skeptical of relying on them (Dawes, 1979). Counter to this notion, results from six experiments show that lay people adhere more to advice when they think it comes from an algorithm than from a person. People showed this sort of algorithm appreciation when making numeric estimates about a visual stimulus (Experiment 1A) and forecasts about the popularity of songs and romantic matches (Experiments 1B and 1C). Yet, researchers predicted the opposite result (Experiment 1D). Algorithm appreciation persisted when advice appeared jointly or separately (Experiment 2). However, algorithm appreciation waned when: people chose between an algorithm’s estimate and their own (versus an external advisor’s; Experiment 3) and they had expertise in forecasting (Experiment 4). Paradoxically, experienced professionals, who make forecasts on a regular basis, relied less on algorithmic advice than lay people did, which hurt their accuracy. These results shed light on the important question of when people rely on algorithmic advice over advice from people and have implications for the use of “big data” and algorithmic advice it generates.

Highlights

Algorithms--scripts for sequences of mathematical calculations or procedural steps--are powerful
The results speak to the strength of reliance on algorithmic advice, especially considering how many decisions are affected by joint-versusseparate evaluation: willingness to pay for consumer goods, willingness to pay for environmental issues, support for social issues, and voter preferences (Hsee, 1996; 1998; Irwin, Slovic, Lichtenstein, & McClelland, 1993; Nowlis & Simonson, 1997)
Did participants rely on the algorithmic advice as much as they should have, given the normative information they received? Compared to the normative benchmark of how much people should have weighted the advice from the person (.5) and algorithm (1), participants underweighted advice from the algorithm more than (M = .66, SD = .34) they did from the person (M = .26, SD = .27), F(1, 669) = 275.08, p < .001, d = 1.30

Summary

Introduction

Algorithms--scripts for sequences of mathematical calculations or procedural steps--are powerful They can complement human judgment and are increasingly used to inform it. When provided the same information, algorithms outperformed expert forecasts of: survival of cancer patients (Einhorn, 1972), severity of pathologies (Goldman et al, 1977), heart attacks (Hedén, Öhlin, Rittner, & Edenbrandt, 1997), recidivism of parolees (Carroll, Wiener, Coates, Galegher, & Alibrio, 1982), magnitude of operational risk (Tazelaar & Snijders, 2013), and answers to trivia questions (Tesauro, Gondek, Lenchner, Fan, & Prager, 2013). This paper examines the apparent distrust in algorithms and simultaneous widespread dependence on them It tests whether people are ever willing to leverage the power of algorithm and if so, when they are most likely to do so.

Methods

Results

Conclusion