Perception of experience influences altruism and perception of agency influences trust in human-machine interactions.

As robots become increasingly integrated into social economic interactions, it becomes crucial to understand how people perceive a robot's mind. It has been argued that minds are perceived along two dimensions: experience, i.e., the ability to feel, and agency, i.e., the ability to act and take responsibility for one's actions. However, the influence of these perceived dimensions on human-machine interactions, particularly those involving altruism and trust, remains unknown. We hypothesize that the perception of experience influences altruism, while the perception of agency influences trust. To test these hypotheses, we pair participants with bot partners in a dictator game (to measure altruism) and a trust game (to measure trust) while varying the bots' perceived experience and agency, either by manipulating the degree to which the bot resembles humans, or by manipulating the description of the bots' ability to feel and exercise self-control. The results demonstrate that the money transferred in the dictator game is influenced by the perceived experience, while the money transferred in the trust game is influenced by the perceived agency, thereby confirming our hypotheses. More broadly, our findings support the specificity of the mind hypothesis: Perceptions of different dimensions of the mind lead to different kinds of social behavior.

Confronting barriers to human-robot cooperation: balancing efficiency and risk in machine behavior.

Many technical and psychological challenges make it difficult to design machines that effectively cooperate with people. To better understand these challenges, we conducted a series of studies investigating human-human, robot-robot, and human-robot cooperation in a strategically rich resource-sharing scenario, which required players to balance efficiency, fairness, and risk. In these studies, both human-human and robot-robot dyads typically learned efficient and risky cooperative solutions when they could communicate. In the absence of communication, robot dyads still often learned the same efficient solution, but human dyads achieved a less efficient (less risky) form of cooperation. This difference in how people and machines treat risk appeared to discourage human-robot cooperation, as human-robot dyads frequently failed to cooperate without communication. These results indicate that machine behavior should better align with human behavior, promoting efficiency while simultaneously considering human tendencies toward risk and fairness.

Cooperating with machines.

Since Alan Turing envisioned artificial intelligence, technical progress has often been measured by the ability to defeat humans in zero-sum encounters (e.g., Chess, Poker, or Go). Less attention has been given to scenarios in which human-machine cooperation is beneficial but non-trivial, such as scenarios in which human and machine preferences are neither fully aligned nor fully in conflict. Cooperation does not require sheer computational power, but instead is facilitated by intuition, cultural norms, emotions, signals, and pre-evolved dispositions. Here, we develop an algorithm that combines a state-of-the-art reinforcement-learning algorithm with mechanisms for signaling. We show that this algorithm can cooperate with people and other algorithms at levels that rival human cooperation in a variety of two-player repeated stochastic games. These results indicate that general human-machine cooperation is achievable using a non-trivial, but ultimately simple, set of algorithmic mechanisms.