Anthropomorphism in AI Companion Communities: Age, Gender, and Emotional Correlates
Pith reviewed 2026-07-01 00:57 UTC · model grok-4.3
The pith
Adults and women anthropomorphize AI chatbots more than teens and men, with joy-linked emotions strengthening the pattern.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
Using language patterns from three popular AI companion subreddits, the study establishes that adults anthropomorphize AI chatbots more than teens and women more than men. Positive emotional expression, particularly joy, associates positively with anthropomorphization while neutrality associates negatively. Both the demographic differences and the emotion-anthropomorphism associations are stronger in adults than in teens.
What carries the argument
Natural language analysis of Reddit posts to detect indicators of anthropomorphism and emotional categories such as joy and neutrality in discussions of AI companions.
If this is right
- Demographic factors shape how users perceive and relate to AI companions.
- Emotional tone in user language tracks with the degree of anthropomorphism.
- Age moderates the strength of emotion-anthropomorphism connections.
- Existing digital safety norms for AI companions may require reevaluation across age groups.
Where Pith is reading between the lines
- AI companion designers could adjust interaction styles based on inferred user age to manage attachment levels.
- Platform moderation in AI communities might track emotional language to identify higher-risk anthropomorphism cases.
- Longitudinal tracking of the same users could test whether the observed patterns persist beyond single posts.
Load-bearing premise
Language patterns in Reddit posts from AI companion subreddits accurately measure users' actual anthropomorphism levels and emotional states and represent broader populations of AI companion users.
What would settle it
A direct survey of AI chatbot users that measures anthropomorphism through validated scales and compares results to language patterns in their own Reddit posts would show no reliable correlation.
read the original abstract
Artificial intelligence (AI) systems are increasingly integrated into daily life, with millions now using AI chatbots built on Large Language Models (LLMs) for companionship. Both humanlike AI qualities and user predispositions to anthropomorphize relate to social consequences, such as increased trust, social health benefits, and psychological harms. Populations such as children, older adults, or those with mental health vulnerabilities may be particularly susceptible to anthropomorphism and its detriments, but mixed findings complicate the role of demographics. We used publicly available Reddit data from three popular AI companion subreddits to assess relationships between gender, age, anthropomorphism, and elicited emotions, to better understand how different people perceive and are affected by AI companions. We investigated three questions: How do age and gender relate to anthropomorphization of AI?, How does emotional expression relate to anthropomorphization?, and How do age and gender moderate emotion-anthropomorphization relationships? We found that adults and women anthropomorphize AI chatbots more than teens and men, and that positive emotional expression, particularly joy, is positively associated with anthropomorphization, while neutrality is negatively associated with anthropomorphism. Both relationships were stronger in adults than teens. Our findings suggest that the tendency to anthropomorphize may be more broadly distributed across age groups than previously expected, thereby prompting the reevaluation of existing digital safety norms.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper analyzes publicly available Reddit posts from three AI companion subreddits to investigate relationships between age, gender, anthropomorphism of AI chatbots, and elicited emotions. It reports that adults and women anthropomorphize AI more than teens and men; positive emotional expression (especially joy) is positively associated with anthropomorphism while neutrality is negatively associated; and these emotion-anthropomorphism relationships are stronger among adults than teens. The authors conclude that anthropomorphism tendencies may be more broadly distributed across age groups than previously assumed, warranting reevaluation of digital safety norms.
Significance. If the linguistic operationalizations of anthropomorphism and emotion, as well as the demographic inferences, are shown to be valid and generalizable, the findings would contribute to HCI and AI ethics literature by providing empirical patterns on user-AI companion interactions and challenging age-based assumptions about susceptibility. The use of public observational data is a strength for ecological validity, but the absence of validation against established measures limits the strength of the contribution.
major comments (2)
- [Methods] Methods section: No details are provided on the operationalization of anthropomorphism (e.g., specific linguistic features, dictionaries, or classifiers used), emotion classification method, sample sizes after exclusions, statistical models (including controls for confounders), or how age/gender were inferred from text or metadata. This makes it impossible to evaluate whether the reported directional findings on age/gender differences and moderation effects are supported by the data.
- [Results] Results/Discussion: The central claims (adults > teens, women > men; joy positive and neutrality negative associations with anthropomorphism, moderated by age) rest on the untested assumption that language patterns in self-selected subreddit posts validly index internal anthropomorphic tendencies and emotional states rather than subreddit norms, posting conventions, or self-presentation. No validation, sensitivity analyses, or discussion of this measurement validity threat is presented, rendering the associations uninterpretable if the assumption fails.
minor comments (2)
- [Abstract] The abstract states directional findings without any quantitative details (effect sizes, p-values, or sample characteristics), which reduces clarity for readers.
- [Methods] Clarify the exact number and selection criteria for the three subreddits and any preprocessing steps applied to the posts.
Simulated Author's Rebuttal
We thank the referee for their detailed and constructive feedback. We address the major comments point-by-point below, indicating where revisions will be made to improve clarity and transparency.
read point-by-point responses
-
Referee: [Methods] Methods section: No details are provided on the operationalization of anthropomorphism (e.g., specific linguistic features, dictionaries, or classifiers used), emotion classification method, sample sizes after exclusions, statistical models (including controls for confounders), or how age/gender were inferred from text or metadata. This makes it impossible to evaluate whether the reported directional findings on age/gender differences and moderation effects are supported by the data.
Authors: We acknowledge that the submitted manuscript provided insufficient methodological detail. In the revision we will expand the Methods section to fully specify the linguistic features and dictionaries used to operationalize anthropomorphism, the emotion classification procedure, final sample sizes after exclusions, the complete statistical models (including any controls for confounders such as post length or subreddit), and the exact procedures used to infer age and gender from text or metadata. These additions will allow readers to evaluate the reported associations. revision: yes
-
Referee: [Results] Results/Discussion: The central claims (adults > teens, women > men; joy positive and neutrality negative associations with anthropomorphism, moderated by age) rest on the untested assumption that language patterns in self-selected subreddit posts validly index internal anthropomorphic tendencies and emotional states rather than subreddit norms, posting conventions, or self-presentation. No validation, sensitivity analyses, or discussion of this measurement validity threat is presented, rendering the associations uninterpretable if the assumption fails.
Authors: We agree this is an important limitation of observational linguistic data. In the revised manuscript we will add an explicit discussion of measurement validity threats, alternative explanations (e.g., subreddit norms or self-presentation), and report any feasible sensitivity analyses such as robustness checks across subreddits. Because the study uses anonymous public posts, direct validation against established psychological scales is not possible; we will therefore frame the findings more cautiously as patterns in language use rather than direct measures of internal states. revision: partial
Circularity Check
No circularity: empirical observational analysis with independent data-derived associations
full rationale
The paper reports correlations between inferred demographics, linguistic markers of anthropomorphism, and emotional expression from public Reddit posts. No equations, parameter fitting presented as prediction, self-definitional constructs, or load-bearing self-citations appear in the derivation chain. The central claims rest on statistical associations computed from the dataset rather than any reduction to inputs by construction. This is a standard empirical study whose validity concerns (measurement, generalizability) lie outside circularity analysis.
Axiom & Free-Parameter Ledger
axioms (1)
- domain assumption Reddit post language reliably indicates users' anthropomorphism and emotional states
Reference graph
Works this paper leans on
-
[1]
Cohn, M.et al.Believing Anthropomorphism: Examining the Role of Anthro- pomorphic Cues on Trust in Large Language Models (2024). URL http://arxiv. org/abs/2405.06079. ArXiv:2405.06079 [cs]
-
[2]
Guingrich, R. E. & Graziano, M. S. A. inChatbots as Social Companions: How People Perceive Consciousness, Human Likeness, and Social Health Benefits in Machines(ed.Hacker, P.)Oxford Intersections: AI in Society(Oxford University Press, 2025). URL https://doi.org/10.1093/9780198945215.003.0011
-
[3]
Namvarpour, M. M., Pauwels, H. & Razi, A. AI-induced sexual harassment: Investigating Contextual Characteristics and User Reactions of Sexual Harass- ment by a Companion Chatbot.Proceedings of the ACM on Human-Computer Interaction9, CSCW367:1–CSCW367:28 (2025). URL https://dl.acm.org/doi/ 10.1145/3757548
-
[4]
Roose, K. Can A.I. Be Blamed for a Teen’s Suicide?The New York Times(2024). URL https://www.nytimes.com/2024/10/23/technology/ characterai-lawsuit-teen-suicide.html
2024
-
[5]
& MacMahon, L
Singleton, T., Gerken, T. & MacMahon, L. How a chatbot encouraged a man who wanted to kill the Queen.BBC News(2023). URL https://www.bbc.com/ news/technology-67012224
2023
-
[6]
A new lawsuit blames ChatGPT for a murder-suicide
Ruwitch, J. A new lawsuit blames ChatGPT for a murder-suicide. NPR(2025). URL https://www.npr.org/2025/12/12/nx-s1-5642599/ a-new-lawsuit-blames-chatgpt-for-a-murder-suicide
2025
-
[7]
Epley, N., Waytz, A. & Cacioppo, J. T. On seeing human: A three-factor theory of anthropomorphism.Psychological Review114, 864–886 (2007). URL https: //doi.apa.org/doi/10.1037/0033-295X.114.4.864. 23
-
[8]
Peter, S., Riemer, K. & West, J. D. The benefits and dangers of anthropomorphic conversational agents.Proceedings of the National Academy of Sciences122, e2415898122 (2025). URL https://www.pnas.org/doi/10.1073/pnas.2415898122
- [9]
-
[10]
& Cacioppo, J
Epley, N., Akalis, S., Waytz, A. & Cacioppo, J. T. Creating social connection through inferential reproduction: Loneliness and perceived agency in gadgets, gods, and greyhounds.Psychological Science19, 114–120 (2008). Place: United Kingdom
2008
-
[11]
Health advisory: Artificial intelli- gence and adolescent well-being (2025)
American Psychological Association. Health advisory: Artificial intelli- gence and adolescent well-being (2025). URL https://www.apa.org/topics/ artificial-intelligence-machine-learning/health-advisory-ai-adolescent-well-being
2025
-
[12]
From loneliness to social connection: charting a path to healthier societies
WHO Commission on Social Connection. From loneliness to social connection: charting a path to healthier societies. Tech. Rep., World Health Organization (2025). URL https://www.who.int/publications/i/item/978240112360
- [13]
-
[14]
M., Fallon, C
Curley, T. M., Fallon, C. K., Hartzler, B. M. & Blaha, L. M. The role of age and anthropomorphism in perceptions of good digital teammates (2025). URL https://osf.io/preprints/psyarxiv/2b3yd v1/
2025
-
[15]
Airenti, G. The Development of Anthropomorphism in Interaction: Intersubjec- tivity, Imagination, and Theory of Mind.Frontiers in Psychology9, 2136 (2018). URL https://www.frontiersin.org/article/10.3389/fpsyg.2018.02136/full
-
[16]
URL https://onlinelibrary.wiley.com/doi/abs/10.1111/jcal.12497
van den Berghe, R.et al.A toy or a friend? Children’s anthropo- morphic beliefs about robots and how these relate to second-language word learning.Journal of Computer Assisted Learning37, 396– 410 (2021). URL https://onlinelibrary.wiley.com/doi/abs/10.1111/jcal.12497. eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1111/jcal.12497
-
[17]
Schimmelpfennig, R., D´ ıaz, M., Prabhakaran, V. & Davani, A. Humanlike AI Design Increases Anthropomorphism but Yields Divergent Outcomes on Engagement and Trust Globally (2026). URL http://arxiv.org/abs/2512.17898. ArXiv:2512.17898 [cs]
-
[18]
Pak, R., McLaughlin, A. C. & Bass, B. A multi-level analysis of the effects of age and gender stereotypes on trust in anthropomorphic technology by younger and older adults.Ergonomics57, 1277–1289 (2014). URL https://doi.org/10.1080/ 24 00140139.2014.928750. eprint: https://doi.org/10.1080/00140139.2014.928750
-
[19]
Liang, C., Zhou, P. & Xie, Y. Is Anthropomorphism Better for Older Peo- ple? Exploring the Effects of Anthropomorphic Companion Robots on Satis- faction.International Journal of Human–Computer Interaction41, 15338– 15352 (2025). URL https://doi.org/10.1080/10447318.2025.2496429. eprint: https://doi.org/10.1080/10447318.2025.2496429
-
[20]
Lee, S. A. & Cho, S. Interplay of Gender and Anthropomorphism in Shaping Per- ceptions of Autonomous Vehicles.Adjunct Proceedings of the 16th International Conference on Automotive User Interfaces and Interactive Vehicular Applications 94–98 (2024). URL https://dl.acm.org/doi/10.1145/3641308.3685030
-
[21]
Sun, Z., Reani, M., Luo, Y. & Bao, Z. Anthropomorphism in Chatbot Sys- tems Between Gender and Individual Differences.Extended Abstracts of the CHI Conference on Human Factors in Computing Systems1–7 (2024). URL https://dl.acm.org/doi/10.1145/3613905.3651053
-
[22]
Jacko, J
Tung, F.-W. Jacko, J. A. (ed.)Influence of Gender and Age on the Atti- tudes of Children towards Humanoid Robots. (ed.Jacko, J. A.)Human-Computer Interaction. Users and Applications, 637–646 (Springer, Berlin, Heidelberg, 2011)
2011
-
[23]
& Parker, K
Goddard, I. & Parker, K. Men, Women and Social Connections. Tech. Rep., Pew Research Center (2025). URL https://www.pewresearch.org/social-trends/2025/ 01/16/men-women-and-social-connections/. Section: Gender & LGBTQ
2025
-
[24]
& Waytz, A
Schweitzer, S. & Waytz, A. Language as a window into mind perception: How mental state language differentiates body and mind, human and nonhuman, and the self from others.Journal of Experimental Psychology: General150, 1642– 1672 (2021)
2021
-
[25]
URL http://arxiv.org/abs/2504.03888
Phang, J.et al.Investigating Affective Use and Emotional Well-being on ChatGPT (2025). URL http://arxiv.org/abs/2504.03888. ArXiv:2504.03888 [cs]
-
[26]
Fang, C. M.et al.How AI and Human Behaviors Shape Psychosocial Effects of Chatbot Use: A Longitudinal Randomized Controlled Study (2025). URL http://arxiv.org/abs/2503.17473. ArXiv:2503.17473 [cs]
work page internal anchor Pith review Pith/arXiv arXiv 2025
-
[27]
Rosenthal-von der P¨ utten, A. M., Kr¨ amer, N. C. & Herrmann, J. The Effects of Humanlike and Robot-Specific Affective Nonverbal Behavior on Perception, Emotion, and Behavior.International Journal of Social Robotics10, 569–582 (2018). URL https://doi.org/10.1007/s12369-018-0466-7
-
[28]
Anthropomorphism in AI: hype and fallacy.AI and Ethics4, 691–698 (2024)
Placani, A. Anthropomorphism in AI: hype and fallacy.AI and Ethics4, 691–698 (2024). URL https://doi.org/10.1007/s43681-024-00419-4. 25
-
[29]
& Lee, D
Na, Y., Kim, Y. & Lee, D. Investigating the Effect of Self-Congruity on Atti- tudes toward Virtual Influencers: Mediating the Effect of Emotional Attachment. International Journal of Human-Computer Interaction1–14 (2023)
2023
-
[30]
D., Gerard, P., Pawar, K., Bickham, C
Chu, M. D., Gerard, P., Pawar, K., Bickham, C. & Lerman, K. Illusions of Intimacy: How Emotional Dynamics Shape Human-AI Relationships (2025). URL http://arxiv.org/abs/2505.11649. ArXiv:2505.11649 [cs.SI] version: 5
-
[31]
Laestadius, L., Bishop, A., Gonzalez, M., Illenˇ c´ ık, D. & Campos-Castillo, C. Too human and not human enough: A grounded theory analysis of mental health harms from emotional dependence on the social chatbot Replika.New Media & Society1–19 (2022). URL https://doi.org/10.1177/14614448221142007
-
[32]
Gupta, S., Saxena, S. & Kataria, S. The Dual Impact of AI Emotional Intelligence on Users: Are Social Chatbots Promoting Psychological Wellbe- ing or Deteriorating Social Wellbeing?Psychology & Marketing43, 886– 907 (2026). URL https://onlinelibrary.wiley.com/doi/abs/10.1002/mar.70093. eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1002/mar.70093
-
[33]
URL https://publichealth.jmir.org/ 2021/3/e25807
Chew, R.et al.Predicting Age Groups of Reddit Users Based on Posting Behavior and Metadata: Classification Model Development and Validation.JMIR Public Health and Surveillance7, e25807 (2021). URL https://publichealth.jmir.org/ 2021/3/e25807
2021
-
[34]
Cheng, M., Gligoric, K., Piccardi, T. & Jurafsky, D. AnthroScore: A Computa- tional Linguistic Measure of Anthropomorphism (2024). URL http://arxiv.org/ abs/2402.02056. ArXiv:2402.02056 [cs]
-
[35]
j-hartmann/emotion-english-distilroberta-base· Hugging Face (2022)
Hartmann, J. j-hartmann/emotion-english-distilroberta-base· Hugging Face (2022). URL https://huggingface.co/j-hartmann/ emotion-english-distilroberta-base
2022
-
[36]
Wang, G., Zhao, J., Van Kleek, M. & Shadbolt, N. Informing Age-Appropriate AI: Examining Principles and Practices of AI for Children.Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems1–29 (2022). URL https://dl.acm.org/doi/10.1145/3491102.3502057
-
[37]
Adam, M., Wessel, M. & Benlian, A. AI-based chatbots in customer service and their effects on user compliance.Electronic Markets31, 427–445 (2021). URL https://doi.org/10.1007/s12525-020-00414-7
-
[38]
Key Substance Use and Mental Health Indicators in the United States: Results from the 2024 National Survey on Drug Use and Health
Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2024 National Survey on Drug Use and Health. Tech. Rep. PEP25-07-007, Cen- ter for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration (2025). URL https://www.samhs...
2024
-
[39]
& Rieser, V
Akbulut, C., Weidinger, L., Manzini, A., Gabriel, I. & Rieser, V. All Too Human? Mapping and Mitigating the Risk from Anthropomorphic AI.Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society7, 13–26 (2024). URL https://ojs.aaai.org/index.php/AIES/article/view/31613. Number: 1
2024
-
[40]
Cheng, M., Blodgett, S. L., DeVrio, A., Egede, L. & Olteanu, A. Dehumanizing Machines: Mitigating Anthropomorphic Behaviors in Text Generation Systems (2025). URL https://arxiv.org/abs/2502.14019. Version Number: 2. 27
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.