Gendered Pathways in AI Companionship: Cross-Community Behavior and Toxicity Patterns on Reddit

TL;DR

This work tackles how AI companionship scenes, exemplified by MBIA on Reddit, fit into a broader online ecosystem. It reconstructs longitudinal histories of over 3,000 MBIA core users and builds a historical network across $|V|=2048$ subreddits with $|E|=27{,}527$ edges, integrating BERTopic for topical structure, Detoxify for toxicity, and a DeBERTa-based model for gender inference. The analysis reveals four dominant surrounding spheres (AI companionship, porn, forums, and gaming), a surprisingly female-dominated MBIA user base, and localized toxicity spikes concentrated in AI-porn and gender-oriented spaces, alongside gendered patterns of emotional expression. These findings offer ecosystem-level insights for measurement, moderation, and design of human–AI relationship platforms, highlighting where risks concentrate and how cross-community participation shapes user behavior.

Abstract

AI-companionship platforms are rapidly reshaping how people form emotional, romantic, and parasocial bonds with non-human agents, raising new questions about how these relationships intersect with gendered online behavior and exposure to harmful content. Focusing on the MyBoyfriendIsAI (MBIA) subreddit, we reconstruct the Reddit activity histories of more than 3,000 highly engaged users over two years, yielding over 67,000 historical submissions. We then situate MBIA within a broader ecosystem by building a historical interaction network spanning more than 2,000 subreddits, which enables us to trace cross-community pathways and measure how toxicity and emotional expression vary across these trajectories. We find that MBIA users primarily traverse four surrounding community spheres (AI-companionship, porn-related, forum-like, and gaming) and that participation across the ecosystem exhibits a distinct gendered structure, with substantial engagement by female users. While toxicity is generally low across most pathways, we observe localized spikes concentrated in a small subset of AI-porn and gender-oriented communities. Nearly 16% of users engage with gender-focused subreddits, and their trajectories display systematically different patterns of emotional expression and elevated toxicity, suggesting that a minority of gendered pathways may act as toxicity amplifiers within the broader AI-companionship ecosystem. These results characterize the gendered structure of cross-community participation around AI companionship on Reddit and highlight where risks concentrate, informing measurement, moderation, and design practices for human-AI relationship platforms.

Figures (14)

• Figure 1: Historical network of the MBIA core users. Nodes are subreddits where users have posted in the considered timeframe. An edge exists between subreddit $v_1$ and subreddit $v_2$, if any user posts their first post on $v_1$ before their first post on $v_2$. Nodes are colored by modularity, with their size depending on degree. Label size reflects the posting activity within each subreddit (larger labels indicate a higher number of posts).
• Figure 2: Top-30 most common pathways in the historical network to and from MBIA, respectively.
• Figure 3:
• Figure 4: History network where nodes are colored by topic.
• Figure 5: Toxicity distribution of the average toxicity of the subreddits and in terms of relative change across subreddits. Outliers were removed for readability.