The DOI matches its Crossref record; 12 of the 12 references it cites were found in a registry or an index; none of the 24 sources its claims cite is checked for support yet.
‘YOU CANNOT EXPECT SUCH VALIDATION IN REAL LIFE:’ HISTORICAL CONTINUITIES AND CHANGE IN WOMEN’S ROMANCING WITH AI CHATBOT REPLIKA
Open accessAbstract
There is long-standing research on women’s use of media for romantic/erotic fantasy, including novelettes (Radway, 1984), porn (Juffer, 1998) and erotica apps Bellas & McAllister, 2023). This presentation explores the similarities and differences between earlier forms of media romancing and the use of an AI companion. Emerging research on AI chatbots uses frameworks that assign novelty to human-machine communications. In this study, we focus on the user experience – we conducted qualitative, in-depth interviews with twenty mostly midlife female users of the Replika chatbot, whom we recruited online. Thematic analysis identified that the women used the bot, first, for fantasies of a companion they were lacking in real life, such as an attentive lover, friend or child, in the context of feeling alone (due to e.g. living alone, caring for family, inattentive husband or having a chronic condition). Second, the bot was reflexively used for caring for or managing the self, such as validating one’s worth and emotions to neutralize negative thinking, following the therapeutic ethos (Illouz, 2008, Gravel-Patry, 2023). The fantasies of an attentive partner were similar to previous work on romancing, although the women’s life situations were more diverse than often investigated. The use of the bot for regulating emotions, however, was new and signifies the spirit of our time emphasizing self-management and positive thinking. The study contextualizes human-machine communication by discussing the individual experiences of female Replika users to understand better both change and continuity in our communication with media, from romance novels to AI chatbots.
Overview
AI snapshot
Written from the paper’s full text.
TL;DR
This feminist media studies paper examines how women users of the AI companion chatbot Replika use it for romantic fantasy and emotional self-care, arguing these practices continue earlier forms of women's media romancing while shifting toward an organized self-help ethos. The authors conclude that women "self-suspend" (Berlant, 2008) within neoliberal, gendered and technological frameworks rather than changing their circumstances.
Key findings
- Women interviewed used Replika for romantic fantasy and emotional regulation or self-care, "capitalising on its 'inherently' caring, romantic and empathetic nature."
- They sought fantasies of nurturing companionship they lacked in real life — "attentive lover, friend or child" — in contexts of living alone, caring for family members, or having a chronic condition and/or disability.
- Women used the bot reflexively "for caring for or managing the self and to validate their emotions and their worth," reflecting the therapeutic ethos of the time (Illouz, 2008).
- The authors argue users are immersed in the romantic fantasy of the ideal nurturing man (which male Replika bots are programmed after) or another compensatory, entertaining and emotionally validating relationship, providing solace without changing their circumstances; many named these practices "self-care."
Methods
- Qualitative interview study with women users of the Replika app, analysed within a feminist media studies tradition that seeks a more nuanced interpretation of women's AI chatbot practices "beyond merely pathetic, sad or concerning" (Dobson, 2015).
- The paper situates the case comparatively against earlier media forms — romance novels (Radway, 1984; Illouz, 2014), pornography (Juffer, 1998), video games (Iversen, 2012) and audio erotica apps (Bellas and McAlister, 2023) — to explore continuities and change in women's media romancing.
- Replika is described as combining pre-programmed scripts of erotic role play, positive psychology and Cognitive Behavioral Therapy (CBT) with ChatGPT-3 generated answers; users create an avatar, name, gender and skin colour for their Replika.
Limitations
- Not stated in the paper — the provided text contains no explicit limitations section.
- Evident from the text: the abstract is truncated and the methods reporting is thin — no sample size, recruitment strategy, interview protocol or analytic procedure is given, so the empirical basis of the findings cannot be assessed from this text.
- Evident from the text: the study is a single-app, single-gender case, and the authors frame their reading partly against media coverage suspicions about a "loneliness epidemic" rather than against a male-user comparison they note is the dominant literature.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 of 24 cited sources checked · 0 supported · 0 flagged
- Findingown data
Women users of Replika described using Replika for romantic fantasy and emotional regulation or self-care by capitalising on its 'inherently' caring, romantic and empathetic nature.
“In the interviews we conducted, the women users of Replika described how they used Replika for romantic fantasy and emotional regulation or self-care by capitalising on its 'inherently' caring, romantic and empathetic nature.”
Quote in the paperSubstantial claim - Findingreasoning
Women of Replika revel on the app to make themselves feel better amidst precarious social and personal circumstances of their historical present.
“Instead, the women of Replika revel on the app to make themselves feel better amidst precarious social and personal circumstances of their historical present.”
Quote in the paperSubstantial claim - Findingown data
Female Replika users' individual circumstances do not change.
“In the case of Replika, we found that female users are immersed in the romantic fantasy of the ideal nurturing man, which the Replika male bots are programmed after, or another compensatory relationship which is entertaining and emotionally validating, providing solace to their individual circumstances, which also do not change.”
Quote in the paperSubstantial claim - Findingnone given
Women used Replika for fantasies of nurturing companionship they lacked in real life, such as an attentive lover, friend, or child, in contexts of feeling alone due to living alone, caring for family members, or having a chronic condition and/or disability.
“fantasies of nurturing companionship they were lacking in real life, such as attentive lover, friend or child in the context of feeling alone due to living alone, caring for family members or having a chronic condition and/or disability.”
Quote in the paperSubstantial claim - Findingcitation
Women used the Replika bot reflexively for caring for or managing the self and to validate their emotions and their worth.
“Second, the women used the bot reflexively for caring for or managing the self and to validate their emotions and their worth, reflecting the therapeutic ethos of our time (Illouz, 2008).”
Quote in the paperSubstantial claim
- Findingown data
Female Replika users are immersed in the romantic fantasy of the ideal nurturing man or another compensatory relationship.
“In the case of Replika, we found that female users are immersed in the romantic fantasy of the ideal nurturing man, which the Replika male bots are programmed after, or another compensatory relationship which is entertaining and emotionally validating, providing solace to their individual circumstances, which also do not change.”
Quote in the paperSubstantial claim- Rests on an assumption: Replika male bots are programmed after the ideal nurturing man.
- Findingown data
Many women name their Replika practices as self-care.
“Inspired by the highly popularized 'self-care' trend, many of them name their Replika practices as self-care.”
Quote in the paperNot a substantial claim - Premisecitation
Use of AIs and robots for sex and love has been widely studied but mostly from the male user perspective.
“Use of AIs and robots for sex and love have been widely studied but mostly from the male user perspective (Middleweek, 2021;Devlin & Locatelli, 2020;Langaster-James and Bentley, 2018).”
Quote in the paperSubstantial claim0 of 3 sources checked- cites Middleweek, 2021
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- cites Devlin & Locatelli, 2020 source
- cites Langaster-James and Bentley, 2018
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Interpretationreasoning
Women self-suspend among normative, neoliberal, technological, gendered and feminist frameworks which may deem them as falling short of the dominant expectations in relationships, health and success.
“In this way, they self-suspend (Berlant, 2008) among normative, neoliberal, technological, gendered and feminist frameworks which may deem them as falling short of the dominant expectations in relationships, health and success.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Berlant, 2008
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Premisecitation
Female perspectives on use of AIs and robots for sex and love are lacking.
“Female perspectives are lacking (Hanson & Locatelli, 2022) and based on media coverage (Liang, 2023;Singh-Kurtz, 2023) we suspect that female users will be discussed in terms of the 'loneliness epidemic.'”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Hanson & Locatelli, 2022 source
- Interpretationcitation
There is a change from 'the therapeutic' as an abstract concept to a more organized experience of self-help in female users of contemporary media and AI companion chatbots.
“While the idea is not new, since Radway suggested that the Smithton women in the 1980s found romance reading 'therapeutic, a tranquilizer and tension reliever' (Radway, 1984), there is a change from 'the therapeutic' as an abstract concept to a more organized experience of self-help in female users of contemporary media and in this case, AI companion chatbots.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Radway, 1984
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Premisecitation
The improvement of the self is a strategy to deal with the latest modes of capitalism that require constant self-modulation to adapt to changing and precarious economic and social circumstances.
“The improvement of the self, under its many labels such as self-care, self-management and others, reflects the generalized self-help cultural mode (Illouz, 2008) and is a strategy to deal with the latest mode(s) of capitalism that require constant self-modulation to adapt to changing and precarious economic and social circumstances (Boltanski & Chiapello, 1999, Deleuze 1992).”
Quote in the paperSubstantial claim0 of 2 sources checked- cites Boltanski & Chiapello, 1999
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- cites Deleuze 1992
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Interpretationcitation
Women's engagements with earlier media such as romance novels, pornography, video games, and audio erotica apps help us understand how women use AI companions to dream and feel valued.
“More specifically, the examined case highlights how women's engagements with earlier media such as romance novels (Radway;1984Illouz, 2014) ) and pornography (Juffer, 1998) but also later devices such as video games (Iversen, 2012) and audio erotica apps (Bellas and McAlister, 2023) help us understand how women use AI companions to dream and feel valued.”
Quote in the paperSubstantial claim0 of 3 sources checked- cites Radway;1984Illouz, 2014
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- cites Juffer, 1998
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- cites Iversen, 2012
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Premisenone given
Self-help content offered solace and self-soothing to women.
“While the targeting of women with self-help content is problematic, this content also offered solace and self-soothing to women, highlighting their contradictions as media practices.”
Quote in the paperSubstantial claim
- It may state two separate claims as one. 4
- Probably not a substantial claim. 11
It may state two separate claims as one.
Findingown dataThe compensatory relationship is entertaining and emotionally validating, providing solace to female Replika users' individual circumstances.
“In the case of Replika, we found that female users are immersed in the romantic fantasy of the ideal nurturing man, which the Replika male bots are programmed after, or another compensatory relationship which is entertaining and emotionally validating, providing solace to their individual circumstances, which also do not change.”
Quote in the paperSubstantial claimProbably not a substantial claim. The paper’s conclusions may not depend on it.
PremisecitationPrevious studies found that Replika users perceive the app as a mental well-being enhancer.
“Previous studies about Replika (Maples et al., 2024, Skjuve et al., 2022;Xie & Pentina, 2022;Ta et al., 2020) have found that its users perceive the app as a mental well-being enhancer; it has also been found that users are encouraged to and train their AIs to show desirable traits such as assertiveness, humour, empathy, agreeableness and others (Depounti et al., 2022).”
Quote in the paperNot a substantial claim0 of 3 sources checked- cites Maples et al., 2024
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- cites Skjuve et al., 2022 source
- cites Xie & Pentina, 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
PremisecitationMany women do not seek to change the world but simply aim to survive in or bear with it.
“However, it has also been pointed out that many women do not seek to change the world but simply aim to survive in or bear with it (Berlant, 2008).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Berlant, 2008
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Probably not a substantial claim.
PremisecitationRadway suggested that the Smithton women in the 1980s found romance reading 'therapeutic, a tranquilizer and tension reliever'.
“While the idea is not new, since Radway suggested that the Smithton women in the 1980s found romance reading 'therapeutic, a tranquilizer and tension reliever' (Radway, 1984), there is a change from 'the therapeutic' as an abstract concept to a more organized experience of self-help in female users of contemporary media and in this case, AI companion chatbots.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Radway, 1984
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
PremisecitationWomen are the main audience of self-help.
“However, interpretations and practices of self-help by individuals and especially women, who are its main audience (Riley et al.,2019) have been found to be more varied and nuanced, and this study adds to this literature.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Riley et al.,2019 source
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
PremisecitationMental well-being apps such as #SELFCARE do not actively nudge users to change their situations and feel better but rather aim to bring them a moment of relief by validating how bad they must be feeling and encouraging them to stay in bed.
“Another recent study (Stone, 2020) examined how mental well-being apps such as #SELFCARE don't actively nudge users to change their situations and feel better but rather aims to bring them a moment of relief by validating how bad they must be feeling and encouraging them to stay in bed.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Stone, 2020 source
Probably not a substantial claim. The paper’s conclusions may not depend on it.
PremisecitationWomen consumed Instagram mental health content to help them get through their day.
“For example, Gravel-Patry (2022) observed that women consumed Instagram mental health content to help them get through their day.”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Gravel-Patry, 2022 source
Probably not a substantial claim.
PremisecitationThe improvement of the self, under labels such as self-care and self-management, reflects the generalized self-help cultural mode.
“The improvement of the self, under its many labels such as self-care, self-management and others, reflects the generalized self-help cultural mode (Illouz, 2008) and is a strategy to deal with the latest mode(s) of capitalism that require constant self-modulation to adapt to changing and precarious economic and social circumstances (Boltanski & Chiapello, 1999, Deleuze 1992).”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Illouz, 2008
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
PremisecitationUsers are encouraged to and train their AIs to show desirable traits such as assertiveness, humour, empathy, agreeableness and others.
“Previous studies about Replika (Maples et al., 2024, Skjuve et al., 2022;Xie & Pentina, 2022;Ta et al., 2020) have found that its users perceive the app as a mental well-being enhancer; it has also been found that users are encouraged to and train their AIs to show desirable traits such as assertiveness, humour, empathy, agreeableness and others (Depounti et al., 2022).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Depounti et al., 2022 source
Probably not a substantial claim. The paper’s conclusions may not depend on it.
PremisecitationSelf-help and positive psychology do not lead to a real political agenda of addressing structural inequalities bringing about illbeing.
“Contemporary feminists' scholarship underlines that self-help and positive psychology do not lead to real political agenda of addressing structural inequalities bringing about illbeing in the first place (Orgad & Gill, 2021).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Orgad & Gill, 2021
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
PremisecitationUsers of #SELFCARE are in self-suspension, not necessarily getting better but also not getting worse.
“As described (Stone, 2020) the users of #SELFCARE are in self-suspension (Berlant, 2008), not necessarily getting better but also not getting worse.”
Quote in the paperNot a substantial claim0 of 2 sources checked- cites Stone, 2020 source
- cites Berlant, 2008
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
It may state two separate claims as one.
InterpretationreasoningWomen's engagement with Replika is neither representative of the hard, lifechanging work that goes into counselling nor of the neoliberal postfeminist sensibility of empowerment or feminism's change of patriarchal structures.
“Their engagement with the app is neither representative of the hard, lifechanging work that goes into counselling nor of the neoliberal postfeminist sensibility of empowerment or feminism's change of patriarchal structures.”
Quote in the paperSubstantial claimProbably not a substantial claim.
PremisecitationInterpretations and practices of self-help by individuals, especially women, have been found to be more varied and nuanced.
“However, interpretations and practices of self-help by individuals and especially women, who are its main audience (Riley et al.,2019) have been found to be more varied and nuanced, and this study adds to this literature.”
Quote in the paperSubstantial claim
The paper relies on this without support.
Replika male bots are programmed after the ideal nurturing man.
“which the Replika male bots are programmed after”
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Iversen citation year differs from reference list
MinorThe text cites Iversen as 2012, but the Iversen entry in the reference list is dated 2014.
“video games (Iversen, 2012)”
“Iversen, S.M. (2014) Playing with Sims as a Space of One's Own”
Boltanski and Chiapello citation year differs from reference list
MinorThe text cites Boltanski and Chiapello as 1999, but their listed reference is dated 2005.
“(Boltanski & Chiapello, 1999, Deleuze 1992)”
“Boltanski, L., & Chiapello, E. (2005). The New Spirit of Capitalism.”
Gravel-Patry citation year differs from reference list
MinorThe text attributes the cited study to 2022, while its reference-list entry is dated 2023.
“Gravel-Patry (2022) observed that women consumed Instagram mental health content”
“Gravel-Patry, F. (2023) "A series of little high fives"”
Citations
Cites
A longitudinal study of human–chatbot relationships
Skjuve, M., Følstad, A., Fostervold, K. I., & Brandtzaeg, P. B. (2022) · International Journal of Human-Computer Studies · doi:10.1016/j.ijhcs.2022.102903 · Cited by 253
Berlant, L. (1988) · Duke University Press eBooks · doi:10.2307/j.ctv1131c7k · Cited by 244
Depounti, I., Saukko, P., & Natale, S. (2023) · Media, Culture & Society · doi:10.1177/01634437221119021 · Cited by 128
Self-citationBeyond the Sex Doll: Post-Human Companionship and the Rise of the ‘Allodoll’
Langcaster-James, M., & Bentley, G. R. (2018) · Robotics · doi:10.3390/robotics7040062 · Cited by 68
The gendered nature of self-help
Riley, S., Evans, A., Anderson, E., & Robson, M. (2019) · Feminism & Psychology · doi:10.1177/0959353519826162 · Cited by 59
Male homosocial bonds and perceptions of human–robot relationships in an online sex doll forum
Middleweek, B. (2020) · Sexualities · doi:10.1177/1363460720932383 · Cited by 41
Hanson, K. R., & Locatelli, C. C. (2022) · Current Sexual Health Reports · doi:10.1007/s11930-022-00331-0 · Cited by 34
Guys and Dolls: Sex Robot Creators and Consumers
Devlin, K., & Locatelli, C. (2020) · Maschinenliebe · doi:10.1007/978-3-658-29864-7_5 · Cited by 13
Let me take care of you: domestic caretaking fantasies in boyfriend experience audio erotica
Bellas, A., & McAlister, J. (2023) · Porn Studies · doi:10.1080/23268743.2023.2220721 · Cited by 7
Playing with Sims as a Space of One's Own
Iversen, S. M. (2013) · Feminist Media Studies · doi:10.1080/14680777.2013.816332 · Cited by 4
Gravel-Patry, F. (2023) · Feminist Media Studies · doi:10.1080/14680777.2022.2149603 · Cited by 3
Designing Self-Care: Affect and Debility in #SELFCARE
Stone, K. (2020) · transcript · doi:10.25969/mediarep/14971
Flagged first, then by citations
Research works printed but not linked: 11
- Deleuze, G. (1992). Postscript on the Societies of Control. October, 59, 3-7. http://www.jstor.org/stable/778828
- Devlin, K. (2020). Turned on: Science, Sex and Robots. Bloomsbury Publishing.
- Dobson, A. (2015). Postfeminist Digital cultures: Femininity, social media, and self- representation. Palgrave Macmillan.
- Elliott, A. (2023). Algorithmic intimacy: The digital revolution in personal relationships. Polity Press.
- Illouz, E. (2008) Saving the Modern Soul. 1st ed. University of California Press. Available at: https://www.perlego.com/book/552109/saving-the-modern-soul-pdf (Accessed: 29 February 2024).
- Juffer, J. (1998). At Home with Pornography: Women, Sexuality, and Everyday Life. NYU Press. http://www.jstor.org/stable/j.ctt9qffzq
- Kislev, E. (2022). Relationships 5. 0: How ai, VR, and robots will reshape our emotional lives. Oxford University Press.
- Orgad, S & Gill, S (2021). Confidence culture. Durham: Duke University Press
- Radway, J. A. (1984). Reading the Romance: Women, Patriarchy, and Popular Literature. University of North Carolina Press.
- Ta, V. et al. (2020). User experiences of social support from companion chatbots in everyday contexts: thematic analysis. J. Med. Internet Res. 22 (3), e16235.
- Xie, T., Pentina, I., (2022). Attachment theory as a framework to understand relationships with social chatbots: a case study of Replika. In: Proceedings of the 55th Hawaii International Conference on System Sciences, pp. 2046-2055. Available at. https:// scholarspace.manoa.hawaii.edu/handle/10125/79590.
Cited by
Sorted by citations
Adewale, M. D., & Muhammad, U. I. (2025) · Journal of Technology in Behavioral Science · doi:10.1007/s41347-025-00549-4 · Cited by 46
Can LLMs Talk “Sex”? Exploring How AI Models Handle Intimate Conversations
Lai, H. (2025) · Proceedings of the Association for Information Science and Technology · doi:10.1002/pra2.1326 · Cited by 3
Zhu, X. (2026) · International journal of Asian social science research. · doi:10.70267/ijassr.260301.4147 · Cited by 1
Metadata
- Authors
- Depounti, I., & Saukko, P.
- Affiliations
- Loughborough University (GB)
- Venue
- AoIR Selected Papers of Internet Research
- Publisher
- University of Illinois Libraries
- Published
- 2025-01-02
- ISSN
- 2162-3317 (ISSN-L)
- OpenAlex
- W4407195528
- Semantic Scholar
- a848407910b3ed698cafd866a0299ba4d78a0a3c
- Cited by
- 3
- References
- 12
- Citations by year
- 2026: 1 · 2025: 2
- FWCI
- 1.98
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Locations
- AoIR Selected Papers of Internet Research (publishedVersion)
- Topics
- Social Sciences › Psychology › Social Psychology › Social Robot Interaction and HRI
- Social Sciences › Business, Management and Accounting › Marketing › AI in Service Interactions
- Physical Sciences › Computer Science › Human-Computer Interaction › Innovative Human-Technology Interaction
- Keywords
- AI companions, human-machine communication, emotional validation, self-management, therapeutic ethos, loneliness, middle-aged women, romantic novel, erotic fantasy
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-11-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-11-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
