Learning with (perceived) humans and computers: the role of agency beliefs in older adult collaborative learning

Learning new skills and information becomes more challenging as people get older, due to a decline in cognitive abilities involved in learning. Learning collaboratively can provide benefits to learning in older age due to the social aspects of collaboration. Prior work has focused on the effects of interacting with another human, while the potential benefits of human-computer interaction are understudied.In the current study, 24 older adult participants (aged 65-80 years) completed an adapted version of the Map Task using a Wizard-of-Oz paradigm, in which participants were told that they were interacting with a human assistant or computer. The findings show that participants became faster across trials in both the Human and Computer conditions, though experienced a steeper reduction in time to complete in the Human learning condition. Despite taking longer, participants did not request additional input from the computer to assist their progress. When recalling their learned routes, participants recalled the route learned with the human more accurately both 1-hour and 1-week after the study session.Overall, participants engaged more with their human partner, and recall was more accurate when the information had been provided by the human, compared to the computer. Participants’ negative perceptions about the computer-provided information and about computers in general may have driven this effect. These findings can inform the future design of learning materials for older adults using natural or synthetic speech, as well as future research on collaborative learning in older age.