Netflix recommendations, Amazon suggestions, Google searches, airline ticket prices, your social media feed. All of these things are driven by algorithms—computer models that crunch massive amounts of data to generate useful results. These types of online algorithms are commonplace and so, generally speaking, we’re used to them. But what about the algorithms behind self-driving cars or airplane autopilots? What about algorithms used to predict crimes or to diagnose medical conditions? These are domains in which it often feels uncomfortable to let a computer model make what could be life-or-death decisions. In this episode of Choiceology with Katy Milkman, we’re exploring the places where algorithms and computer models bump up against resistance from their human users. Seeing as it’s Super Bowl season, it seemed like a good time to revisit last year’s contest as a case study in decision making. The 2018 Super Bowl champion Philadelphia Eagles played incredibly well against the formidable New England Patriots. The game could have gone either way, but the Eagles had a secret weapon that gave them an advantage. We speak with Michael Kist from Bleeding Green Nation on the Eagles’ integration of computer models for decision making both on and off the field. You’ll hear the story of how those models were temporarily abandoned and the team struggled before re-embracing them. Next, we explore the way self-driving cars make split-second decisions on the road, with results that can make their human passengers squirm. We test whether or not giving people a small amount of control over how a self-driving car behaves gives those people a bit more confidence about the technology. Then Katy speaks with her Wharton School of Business colleague Cade Massey, who explains some of the fascinating ways that algorithms have improved decision making and looks at some of the scenarios where algorithms face an uphill battle for acceptance. Cade Massey is a partner in Massey-Peabody Analytics. Finally, Katy recaps the ways that people designing—or simply using—algorithms can work to overcome our human tendency toward machine mistrust.