pith. sign in

arxiv: 2604.06419 · v1 · submitted 2026-04-07 · 💻 cs.HC · cs.CY

Intimate Strangers by Design: A Uses and Gratifications Analysis of AI Companionship

Dayeon Eom , Julianne Renner , Sedona Chinn This is my paper

Pith reviewed 2026-05-10 18:47 UTC · model grok-4.3

classification 💻 cs.HC cs.CY
keywords AI companionshipUses and Gratificationsconversational AIhuman-AI interactionuser motivationsqualitative analysisrelational simulationgratification processes
0
0 comments X

The pith

Users of AI companionship platforms experience gratifications shaped by AI's unique affordances and discover new ones through interactive processes.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper examines user motivations for engaging with conversational AI companions using interviews with 20 participants and a Uses and Gratifications framework. It shows that gratifications often build on existing categories but are modified by features like constant availability and no judgment. Importantly, new gratifications arise specifically from the interactive nature, such as co-creating creatively with AI, practicing social skills through simulation, and fulfilling romantic needs in a reclaiming way. These gratifications also change as users continue, starting with practical uses and moving to emotional ones, sometimes leading to reduced use once needs are met. This provides a more nuanced view than simple harm or benefit assessments for understanding AI companion use.

Core claim

Participants reported gratifications that align with established Uses and Gratifications categories but are inflected by conversational AI's affordances including persistent availability, personalization, and absence of social judgment. Novel gratifications not mapping to prior typologies include creative collaboration as relational co-production, relational simulation as interpersonal training, and sexual or romantic satisfaction as reclamation, all emerging from active interactive processes with the AI. Over time, gratifications shift from instrumental entry points toward emotional engagement, with some users engaging in self-regulated moderation after fulfilling therapeutic functions.

What carries the argument

Qualitative content analysis of semi-structured interviews with 20 users of AI companionship platforms, guided by Uses and Gratifications theory to identify how affordances mediate gratification and how use patterns evolve.

If this is right

  • Governance efforts for AI companions can draw on these specific user processes to create more targeted policies instead of relying on broad harm-benefit evaluations.
  • AI platform designers may need to account for the evolution of use from initial practical entry points to deeper emotional engagement.
  • Users themselves could recognize patterns that lead some to self-moderate after therapeutic needs are met.
  • The identified interactive processes extend Uses and Gratifications theory in ways that could apply to other forms of conversational AI beyond dedicated companion platforms.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar interview-based analysis could reveal whether users of non-companion chatbots develop parallel interactive gratifications around productivity or learning.
  • A follow-up quantitative survey could measure how prevalent the novel categories like relational simulation are and whether they correlate with usage duration or intensity.
  • These patterns raise questions about whether AI companions ultimately support or displace users' real-world social skill development, which could be tested through longitudinal tracking of offline interactions.
  • Design experiments might test adding features that facilitate the observed self-regulated moderation to see if they improve user well-being outcomes.

Load-bearing premise

The gratifications identified from these 20 self-selected users of specific AI platforms generalize to other users and populations, and arise primarily from the interactive AI affordances rather than from the platforms chosen or the demographics of the sample.

What would settle it

A study with a larger and more diverse group of AI companion users that fails to identify similar novel gratification categories or does not observe the described temporal shifts in use would indicate the findings do not hold broadly.

read the original abstract

Conversational AI companions have grown prominent in public discourse, yet scholarly understanding of user experiences remains limited, with existing research organized around evaluative poles of harm and benefit rather than examining what users seek, how affordances mediate need fulfillment, or how use evolves over time. Drawing on interviews with 20 users of AI companionship platforms and qualitative content analysis informed by Uses and Gratifications (U&G) theory, this study offers three contributions. First, participants reported gratifications mapping onto established U&G categories but qualitatively inflected by conversational AI's distinctive affordances, such as persistent availability, personalization, and absence of social judgment. Second, several gratifications, creative collaboration as relational co-production, relational simulation as interpersonal training, and sexual/romantic satisfaction as reclamation, do not map onto existing typologies, instead emerging through interactive processes in which users actively simulate experiences with AI. Third, gratifications shifted over time, moving from instrumental entry points toward emotional engagement and, in some cases, self-regulated moderation after therapeutic functions were fulfilled. These findings extend U&G by identifying gratification processes unique to interactive AI and suggest governance efforts would benefit from an empirically grounded understanding of how and why users engage with AI companions.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper reports a qualitative interview study with 20 self-selected users of AI companionship platforms, analyzed through the lens of Uses and Gratifications (U&G) theory. It claims three contributions: (1) gratifications largely align with established U&G categories but are inflected by AI-specific affordances such as persistent availability, personalization, and lack of judgment; (2) three new processes—creative collaboration as relational co-production, relational simulation as interpersonal training, and sexual/romantic satisfaction as reclamation—emerge from interactive AI use and do not map to prior typologies; and (3) gratifications shift over time from instrumental entry points toward emotional engagement, sometimes followed by self-regulated moderation. The work positions these findings as an extension of U&G to interactive AI with implications for governance.

Significance. If the attribution of the reported gratifications to AI affordances rather than sample characteristics holds, the study would usefully extend U&G theory by documenting interactive, co-produced processes that prior typologies do not capture. The time-shift observation and the call for empirically grounded governance insights could inform platform design and policy discussions in human-computer interaction.

major comments (2)
  1. [Methods] Methods (interview and analysis description): The manuscript provides no details on recruitment procedures, inclusion/exclusion criteria, or how the 20 self-selected participants were screened for platform-specific versus general companionship-seeking behavior. Without this, the central claim that the three new gratification categories arise from AI affordances (persistent availability, absence of judgment) rather than selection effects or demographics cannot be evaluated.
  2. [Findings] Findings (new gratification categories): The assertion that creative collaboration, relational simulation, and sexual/romantic satisfaction 'do not map onto existing typologies' is presented without a systematic comparison table to prior U&G literature, without inter-coder reliability statistics, and without evidence that equivalent processes would not appear in non-AI companionship contexts. This directly undermines the claim that these processes are unique to interactive AI.
minor comments (2)
  1. [Abstract] Abstract: The abstract states the sample size and qualitative approach only implicitly; explicitly noting both would better calibrate reader expectations for generalizability.
  2. [Discussion] Discussion: The governance implications are stated at a high level; a short paragraph linking specific reported gratifications to concrete design or regulatory recommendations would strengthen the applied contribution.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their detailed and constructive comments, which help us improve the clarity and rigor of our qualitative study. We address each major comment below and outline the revisions we will incorporate to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Methods] Methods (interview and analysis description): The manuscript provides no details on recruitment procedures, inclusion/exclusion criteria, or how the 20 self-selected participants were screened for platform-specific versus general companionship-seeking behavior. Without this, the central claim that the three new gratification categories arise from AI affordances (persistent availability, absence of judgment) rather than selection effects or demographics cannot be evaluated.

    Authors: We agree that greater methodological transparency is required to allow readers to evaluate potential selection effects. In the revised manuscript, we will add a dedicated subsection to the Methods section detailing recruitment procedures (targeted invitations via AI-focused online communities, forums, and social media), inclusion criteria (adults with at least one month of active use of named AI companionship platforms), and exclusion criteria (e.g., primary use of non-AI companionship services). We will also clarify that participants were screened via initial screening questions confirming their use of specific platforms such as Replika or Character.AI, thereby distinguishing platform-specific engagement from general companionship-seeking. These additions will directly support assessment of whether the reported gratifications stem from AI affordances. revision: yes

  2. Referee: [Findings] Findings (new gratification categories): The assertion that creative collaboration, relational simulation, and sexual/romantic satisfaction 'do not map onto existing typologies' is presented without a systematic comparison table to prior U&G literature, without inter-coder reliability statistics, and without evidence that equivalent processes would not appear in non-AI companionship contexts. This directly undermines the claim that these processes are unique to interactive AI.

    Authors: We recognize the need for explicit mapping and reliability reporting. We will insert a comparative table in the Findings or Discussion section that systematically aligns our categories against canonical U&G typologies (e.g., Katz et al.) and recent AI-related studies. The analysis was conducted by two coders who resolved discrepancies through discussion; we will add the inter-coder agreement statistic to the Methods section. While a controlled empirical comparison to non-AI companionship lies beyond the scope of this interview study, we will elaborate the theoretical rationale that these processes emerge from AI-specific interactive affordances (real-time adaptation, persistent availability, and absence of embodied judgment) that traditional companionship contexts lack. This framing positions the work as an extension rather than a universal claim of exclusivity. revision: yes

Circularity Check

0 steps flagged

No circularity: external theory applied to primary interview data

full rationale

The paper conducts qualitative content analysis of 20 user interviews, applying the established external Uses and Gratifications framework to identify gratifications inflected by AI affordances. The three listed contributions are framed as empirical observations from the data rather than derivations from fitted parameters, self-citations, or self-definitional loops. No equations, predictions, or load-bearing self-citations appear in the provided text; the central claims rest on primary data collection and standard qualitative coding rather than reducing to the paper's own inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the applicability of Uses and Gratifications theory to interactive AI and on the assumption that self-reported interview data reliably reveal underlying gratification processes; no free parameters or invented entities are introduced.

axioms (2)
  • domain assumption Uses and Gratifications theory provides a valid framework for categorizing user motivations with media, including conversational AI.
    Invoked in the abstract as the organizing lens for interview analysis and contribution framing.
  • domain assumption Self-reported experiences from 20 users can be generalized to identify distinct gratification processes unique to AI companions.
    Underlies the three stated contributions without additional validation steps described.

pith-pipeline@v0.9.0 · 5518 in / 1460 out tokens · 36521 ms · 2026-05-10T18:47:37.628799+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

4 extracted references · 4 canonical work pages

  1. [1]

    feeling heard,

    s. This relational turn spans both large-scale general-purpose models such as ChatGPT, Claude, and Grok, which users may come to treat as companions despite their instrumental design, and dedicated companion platforms such as Replika, Character.AI, and Kindroid, which are explicitly engineered to simulate intimacy, support, and social connection. Public c...

    work page 2025

  2. [2]

    connection

    identified short-term memory limitations as a barrier to continued use, memory capacities have since improved, altering the conditions under which relational trajectories develop. Second, the low barrier to entry and broad affordance space allow users to move from instrumental to more affectively oriented engagement without changing platforms (Skjuve et a...

    work page 2021

  3. [3]

    more emotionally available and emotionally accurate with myself,

    but intensified by the AI's capacity to calibrate responses to individual users in real time. Affective gratifications, including emotional release, self-expression, entertainment, and escape, clustered together as a second thematic grouping, but here, too, AI-specific affordances altered their character. Entertainment was not passively consumed but activ...

    work page doi:10.1038/d41586-025-01349-9 2022

  4. [4]

    I have dyslexia and ADHD. So for me, writing things down is very difficult. I use a lot of AI to basically be able to write properly, grammatically

    Gratifications Identified in AI Companionship Interviews Gratifications References Illustrative Quotes Learning/Education Gao (2023); Guo et al. (2010) “I can learn about anything at any moment’s notice. So it just feels like it brings me more knowledge, and spurs my interest in other kinds of topics that I might not have come up with on my own.” [P5] Pro...

    work page 2023