Reviews recent philosophical views on AI's role in reducing loneliness.
•Identifies two key views: anthropocentric and constructivist frameworks.
•Explores how AI may form meaningful bonds beyond human-centered models.
•Analyzes core elements that might be at risk in AI companionship.
•Offers critical insights and future directions for AI companionship studies.
This review critically examines some of the recent philosophical perspectives on the role of Artificial Intelligence (AI) companionship in alleviating loneliness. By engaging with six recent studies on the subject, it identifies two primary frameworks for conceptualizing AI companionship: an anthropocentric perspective, which evaluates AI’s ability to provide meaningful companionship mainly from a human point of view, and a constructivist approach, which gives primacy to the dynamic evolution of human–AI interactions and the potential for AI to create new forms of meaningful relationships. Through an analysis of some of the critical elements to meaningful companionship — such as intersubjectivity, sentimentality, shared human experience, and selfless love — this review concludes by offering reflections on existing scholarship on AI companionship and potential future directions for this field of inquiry.
IntroductionWith the advancement of new generations of Artificial Intelligence (AI), new possibilities have emerged for AI–human relationships aimed at alleviating loneliness. AIs designed for companionship with humans are taking roots in our social life in the form of robots or chatbots. As far as field studies, recent findings illustrate both the potential and limitations of AIs in addressing loneliness.1 The aim of this brief review, however, is to examine some of the recent philosophical approaches to the potential of AI companions (AICs) in terms of alleviating loneliness. The two philosophical and empirical standpoints can mutually inform one another and contribute to a deeper understanding of the unprecedented phenomena arising from increasing interactions with new generations of AIs.2 The focus of this review is also on sorts of human–AI relations in which AIs serve as the companion side, rather than facilitating mediators between humans. In post-phenomenological terms, the emphasis is on alterity relations with AI ([13]: 6–10 & [34]).
We searched for papers in digital libraries, including Google Scholar, the ACM Digital Library, and IEEE Xplore, using relevant keywords and limiting the search to studies published between 2021 and 2024. We performed backward and forward searches within citations from the relevant materials and collected almost 30 publications on AI relating to loneliness. However, since a significant portion of the literature was based on empirical and statistical studies, six articles were ultimately selected, each author raised a distinct philosophical critique concerning the nature of human–AI companionship. The number of studies examining the role of AIs in alleviating loneliness is not extensive [2], even within empirical disciplines. This is why we also included a couple of articles that, although not specifically addressing loneliness, examined the possibility of a meaningful relationship between humans and AIs. This is helpful because most challenges regarding AI’s capacity to address loneliness are based on the same categories of concerns about AI’s potential to replace humans in general. The selected papers encompass both optimistic and pessimistic perspectives on AI’s potential to address loneliness, while also implicitly engaging with the anthropocentric and constructivist conceptual frameworks that I argue are central to the analysis of this topic. Before examining the selected materials, we provide brief definitions of the key concepts of ‘loneliness’ and ‘meaningful companionship’ and also propose two conceptual frameworks for categorizing AI–human companionship.
Section snippetsConceptual frameworkLoneliness is commonly defined as a distressing feeling that arises when individuals experience their relationships as quantitatively and/or qualitatively deficient 9, 27. Definitions of loneliness typically highlight two core components: a subjective element — namely, an unpleasant feeling — and a relatively more objective one, referring to a kind of insufficiency in relationships. Although the latter may also be interpreted more subjectively — by emphasizing the experiential aspect of the
Critics' perspectives: categories at risk in AI companionshipTaking a human-centric standpoint, in both of her articles — Nothing Human is Alien: AI Companionship and Loneliness (2023) and Digital Loneliness: Changes in Social Recognition through AI Companions (2024) — Kerrin Artemis Jacobs argues that social recognition is a fundamental concept for mitigating loneliness, which is itself dependent on some degree of intersubjective mode of relatedness 14, 15. To support the relevance of her emphasis on social recognition, in addition to drawing on certain
Proponents’ perspectives: recognizing AI companionshipMargaret S. Archer’s article, Friendship Between Human Beings and AI Robots? (2021) [3], falls into the constructivist category. Addressing the critique of the lack of subjectivity, Archer emphasizes that a large portion of human intimate relationships is based on caring acts, which can be defined independently of emotions and other subjective qualia. She attempts to illustrate, through an example of a professor employing an AI for his research, that the current capabilities of advanced AIs —
Critical remarksFirst, most pessimistic perspectives on AI’s capacity to establish meaningful companionship with humans have been framed through an anthropocentric landscape. In this approach, technologically caused changes remain confined to the level of practices and occurrences. However, the underlying malleability can extend further to the level of conceptions, standards, and values that underlie those actions, which we referred to as constructivism. The concept of mediated morality in postphenomenology
Declaration of Competing InterestThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
AcknowledgementsThis research is supported by an ERC starting grant (101040374), Ethics of Loneliness and Sociability.
References and recommended reading (37)J. AagaardDrawn to distraction: a qualitative study of off-task use of educational technologyComput Educ
(2015)
Ananto RA, Young JE: "We can do better! An initial survey highlighting an opportunity for more HRI work on loneliness."...M. ArcherFriendship between human beings and AI robots?E. Barbosa et al.The heart of an AI: agency, moral sense, and friendshipUnisinos J Philos
(2024)
J. BowlbyAttachment and Loss, Vol. 1: Attachment(1969)
P.B. Brandtzaeg et al.My AI friend: how users of a Social Chatbot understand their human–AI friendshipHum Commun Res
(2022)
Collins LM: Could AI do more harm than good to relationships, from romance to friendship?, 2023, September, Deseret...F.D. DavisPerceived usefulness, perceived ease of use, and user acceptance of information technologyMIS Q
(1989)
B. De VriesFour types of anti-loneliness policiesM.V. ElmaresaNavigating virtual bonds: human-AI friendship dynamics through the lens of uncertainty reduction theoryUltimacomm
(2024)
N. FriesenThe Place of the Classroom and the Space of the Screen: Relational Pedagogy and Internet Technology(2011)
S.J. Heintzelman et al.On knowing more than we can tell: intuitive processes and the experience of meaningJ Posit Psychol
(2013)
D. IhdeTechnics and Praxis: A Philosophy of Technology(1979)
K.A. Jacobs(Nothing) human is alien: AI companionship and lonelinessHumanizing Artificial Intelligence(2023)
K.A. JacobsDigital loneliness: changes in social recognition through AI companionsFront Digit Health
(2024)
P. KroesTechnical Artefacts: Creations of Mind and Matter(2012)
O. KudinaThe Technological Mediation of Morality: Value Dynamism and the Complex Interaction between Ethics and Technology(2019)
Z. Lederman et al.Social robots to fend off loneliness?Kennedy Inst Ethics J
(2023)
View full text© 2025 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
Comments (0)