From Extended Self to Posthuman Other: Empathy and Morality in Encounters with AI Characters

This paper examines how contemporary videogames use empathy to shape players' moral perceptions of artificial intelligence, with particular attention to non-playable characters rather than player avatars. The authors frame empathy as affective or cognitive perspective-taking and morality as intuitive judgments of right and wrong, situating their inquiry within an ongoing debate over whether empathy is inherently prosocial or can instead encourage identification with morally flawed figures. They also note longstanding criticism of games and virtual reality as "empathy machines" that may commodify marginalized experiences.

The authors argue that the question of AI's moral status is entangled with its perceived capacity for empathy. Research in psychology suggests people attribute agency and moral responsibility to AI and feel empathy toward it, even where systems lack genuine emotional experience. The paper treats these gaps between cultural imaginaries and technical capabilities not as mere deficits in AI literacy but as sites where society negotiates values and ethics for the digital age, with fictional media serving as key arenas for that negotiation.

The study proposes a taxonomy of three configurations of AI non-playable characters: companions, which typically invite other-oriented empathy and reassure players' moral intuitions; opponents and overlords, which often provoke frustrating forms of empathy tinged with pity, guilt, or shame; and posthuman Others, which pose demanding questions about human-nonhuman kinship and subjectivity. Drawing on concepts such as difficult and hard empathy, the authors conduct close readings of a heterogeneous corpus of games, including Halo, Portal, Nier: Automata, Cyberpunk 2077, Norco, and Citizen Sleeper. The work is presented as a preliminary, interdisciplinary anthological study rather than a quantitative survey, and it does not report sample sizes or empirical player data.