The Use of Robotic Players in Online Games

Short interactive games allow students to gain first-hand experience of incentives and their
impact on decision-making. However, the move to an online teaching environment due to
Covid-19 presented challenges for the use of games, as they typically require synchronous
human-human interaction.
Consequently, the widespread adoption of asynchronous activities means that students
cannot play such interactive games against one another. An alternative is to run them
where students play against robotic players that make decisions according to some preprogrammed rules. The aim of our study is to investigate how student engagement and
in-game behaviour changes when robotic players are used.
We used the webinars in a core first-year microeconomics module to play the Prisoner’s
Dilemma game. In the version of the game we played, students were paired with the same
player for several rounds. This environment creates the classic tension between cheating
and cooperating: joint pay-offs are maximised through cooperation, but each player has an
incentive to cheat in any given round. We used the webinars to implement four different
treatments: students were either knowingly or unknowingly paired with other students
or robotic players. This method allowed us to identify any differences in the propensity
to cooperate. Participants were also required to complete questionnaires about their
perceptions of the usefulness of the game.
With respect to student engagement, our results show that the perceptions of the game
were similar across all treatment groups. However, in-game behaviour did change. Using
a Probit model, we found that students were less likely to cooperate when knowingly
playing against a robotic player. In particular, students were more likely to cheat when
cooperation had been established in the previous round.