Scammed By Robots: Can Deception by Two Simple Non-Humanoid Robots Shape General Attitudes toward Robots?

Abstract

As multi-robot environments become more common, it is important to assess the robots' impact on the humans interacting with them and its magnitude. We investigated whether an interaction with two simple non-humanoid robots can be perceived as an intentional collusion aimed to deceive the person interacting with them. We evaluated if such a negative interaction would result in shaping people's perception of the robots in the interaction and extend to shaping their attitudes towards robots in general. In the study, participants played a gambling game with two simple non-humanoid robots: a “Dealer” responsible for card shuffling and an “Advisor” providing recommendations. We tested participants' experience when the “Advisor” robot gave a false recommendation in the final round, causing them to lose all their winnings. Our findings indicated that this single instance of perceived deception led to a negative perception of the “Advisor” robot, and also negatively influenced people's attitudes toward robots in general, at least temporarily. This study highlights the importance of carefully designing interactions in multi-robot environments, even when the interaction involves very simple robots. Careful consideration of ingroup-outgroup effects and verifying that the robots' collaboration is not interpreted as intentional deception is essential for avoiding powerful negative effects that reach beyond the specific interaction with the robots.