Can humans be friends with social robots?

In this thesis, I will argue for three central claims. Firstly [Chapters Two-Five], on my neo-Aristotelian account of friendship, I argue that humans and social robots that currently exist are not engaged in a relationship of friendship. This is because human-robot interaction (HRI) cannot satisfy two conditions that are necessary for friendship: ‘The Mutual Love Condition’ and ‘The Mutual Recognition Condition.’ My second central claim [Chapter Six and Seven] is that some types of HRI are characterisable as quasi-friendships. By ‘quasi-’ I mean ‘apparently but not really.’ These relationships are apparently friendships because humans can imagine that they are friends with a robot (i.e., that the necessary conditions of friendship are met) when they are not. Moreover, these relationships are apparently friendships because humans and robots can manifest, or imitate manifesting, contingent properties that are common to real friendship. For example, gifting, spending time together, and exchanging words of affirmation. This account of quasi-friendship explains the resemblance between human-robot interaction and friendship, and the ethical significance of quasi-friendship, without conceding that humans and robots are in fact friends. My third central claim [Chapter Eight] is that one can demarcate cases where a user is delusional when they are forming quasi-friendships with robots, from cases where they are playing a game, in ways that are ethically significant. I argue that if some humans have a (certain kind of) belief that they are friends with a robot, then they are delusional. By contrast, I argue that other users who do not believe that they are friends with a robot, but who nonetheless imagine that they are friends with a robot (e.g., because they ‘make-believe’ that they are) can be conceptualised as playing an imaginative game of friendship(q) with the robot. I highlight the distinct ethical concerns relevant to each case. Ultimately, I will argue that in either case, human-robot quasi-friendships are not inherently ethically bad and may even be of significant value to some humans. However, I argue that human-robot quasi-friendships are not as valuable as human-human friendships, and they are not welcome in the ideal human life.