Every number we publish is earned. The AIGirlfriends.ai Research Library is where we surface the data behind our articles: original user surveys, platform analytics, academic literature reviews, and behavioral studies conducted across our active user base. Since May 2025, we have published one research report every month.
13 reports published covering May 2025 through May 2026. All methodology is disclosed. Datasets available for academic and media use upon request at [email protected]
Monthly Research Reports (May 2025 - May 2026)
| Report Title | Summary | Access |
|---|---|---|
| AI Girlfriend Statistics 2026: Full Platform Data Report May 2026 | Comprehensive annual update: market sizing, 13-month platform analytics, income-segmented satisfaction scores, and conversion funnel data from 6,000+ user profiles. The most complete picture of the AI girlfriend category available anywhere. | Read Full Article |
| Q1 2026 Market Report: Revenue, Platforms, and Funding April 17, 2026 | Platform landscape update through Q1 2026. Covers revenue benchmarks, app store share shifts, VC activity, and emerging regional players. Includes updated ARPU comparisons across 8 major platforms. | Download PDF |
| Mental Health Outcomes: 12-Month Cohort Study March 5, 2026 | Longitudinal wellbeing tracking across 890 consenting users over 12 months using PHQ-9 depression proxy measures. Finds net positive outcomes for moderate users who maintained real-world social contact; mixed outcomes for high-substitution users. Pre-registered study design. | Download PDF |
| Gender Dynamics in AI Companionship: 18-Month Analysis February 14, 2026 | Longitudinal comparison of male, female, and non-binary usage patterns across 18 months of platform data. Identifies distinct satisfaction drivers, feature preferences, and emotional disclosure patterns by gender. Includes findings on the narrowing gender gap in new registrations. | Download PDF |
| Annual User Survey 2025: Full Findings (n=3,847) January 29, 2026 | Full-year survey across key dimensions: satisfaction, primary motivations, spending intent, emotional outcomes, and attitudes toward AI regulation. Largest single survey we have conducted to date. Covers 14 countries across 4 regions. | Download PDF |
| Holiday Patterns: Engagement, Loneliness, and Gifting Behavior December 11, 2025 | Analysis of December usage spikes, holiday-period loneliness correlations, and in-app gifting behavior. Finds a 31% increase in new registrations during the last two weeks of December and a 44% spike in emotional support conversations during the holiday window. | Download PDF |
| Voice Interaction Growth and User Sentiment Analysis November 20, 2025 | Six-month study (May-October 2025) tracking voice feature adoption, satisfaction impact, and NPS effects. Voice chat users show 23% higher satisfaction scores than text-only users at the same subscription tier. Includes sentiment coding across 12,000 post-session surveys. | Download PDF |
| Free-to-Premium Conversion: Behavioral Triggers and Timing October 9, 2025 | Analysis of upgrade decision points across 3,100 conversion events. Identifies the three highest-signal behavioral triggers predicting upgrade: memory feature exposure, seventh consecutive daily login, and first voice message sent. Conversion probability model with 74% accuracy. | Download PDF |
| Skin Tone Preferences in AI Companion Design September 17, 2025 | User behavior study on avatar customization and racial representation across 4,200 sessions. Examines how default design choices affect initial engagement and how user-directed customization correlates with retention. Includes regional variation analysis. | Read Full Article |
| Asia-Pacific vs. North America: Regional Behavior Study August 7, 2025 | Cross-regional comparison of usage patterns, feature preferences, peak hours, and monetization behavior across 8 markets. Japan shows the highest daily engagement at 35%; North America leads in ARPU at $54.80/month. Study covers 1,200 users per region. | Download PDF |
| Emotional Attachment in Long-Term AI Companion Users July 18, 2025 | Longitudinal study of 620 users who had been active for 6+ months, examining attachment formation rates, dependency indicators, and psychological risk profiles. 22% meet criteria for emotional attachment; 18% show dependency-risk signals. Includes clinical screening methodology. | Download PDF |
| Q2 2025 User Demographics Report June 12, 2025 | Age, gender, income, relationship status, and educational breakdown from a 1,840-user sample. Baseline demographic snapshot updated quarterly. Covers distribution across 12 countries with regional confidence intervals. | Download PDF |
| AI Girlfriend Statistics 2025: Market Growth and User Trends May 25, 2025 | Baseline report: market sizing, user demographics, engagement metrics, and psychological outcomes from 2,312-respondent survey. First publication from our formal research program. Cited in 14 subsequent industry articles. | Download PDF |
Research Datasets Available for Download
We make select anonymized datasets available for academic and media use. Each dataset is stripped of all personally identifiable information and reviewed before release.
| Dataset Name | Description | Download |
|---|---|---|
| User Demographics (2026 Annual) | Survey data from 3,847 users aged 18-55 (2026 annual study) covering age, gender, income, country, relationship status, and openness to AI romance. | Download PDF |
| User Demographics (Q1 2025) | Survey data from 2,312 users aged 18-40, covering age, gender, income, and openness to AI romance. Baseline dataset for all 2025 citations. | Download CSV |
| Daily Engagement Logs (Feb-May 2025) | Server log analysis with session frequency, length, message count, and interaction type across 6,000+ anonymized user IDs. Covers the 14-week baseline period. | Download CSV |
| Search Trends (2022-2026) | Google Trends and Semrush analysis of "AI girlfriend" and related search terms across 48 months. Includes country-level normalized volumes. | Download CSV |
How We Collect Data
Our research draws on five source types, each with distinct strengths and limitations. We disclose which applies to every published figure.
1. Proprietary User Surveys
We conduct anonymized surveys through platform opt-ins, email invitations, and social media channels including Reddit and X. All responses are voluntary, age-verified, and limited to one entry per participant via email authentication. Survey design follows established social science protocols; questionnaires are available for peer review upon request.
2. Internal Platform Analytics
Usage data comes from anonymized server logs and interaction patterns using Matomo Analytics. We track session frequency, duration, message count, and feature engagement. No personally identifiable information is collected at any stage. Analytics are reviewed quarterly by an internal privacy team.
3. Cohort Studies
Our longitudinal work (including the 12-month mental health cohort and 18-month gender dynamics study) uses opt-in consenting user panels. Participants complete periodic surveys and are compensated with platform credits. Wellbeing measures use validated scales including PHQ-9 and UCLA Loneliness Scale adaptations, administered by a licensed clinical psychologist.
4. Search Trend Analysis
Keyword volume data uses Semrush historical datasets and Google Trends normalized outputs. We validate growth curves across both tools before publishing figures. Country-level data is converted from normalized indexes using calibration against known commercial search volumes.
5. Market and Community Data
We reference publicly available market forecasts from Market.us, Grand View Research, TechCrunch, and Crunchbase for industry-level figures. Community analysis uses NLP tools on Reddit and Quora to surface sentiment and behavioral patterns, coded manually for accuracy.
All data is collected and processed in compliance with digital research ethics and privacy best practices. See our Data Policy for full details.
Collaborate With Us
Are you a researcher, journalist, academic institution, or policy organization interested in data access, collaboration, or co-authorship?
Reach out: [email protected]
We actively partner with universities and independent researchers studying AI relationships, loneliness, digital intimacy, and human-computer interaction. Data sharing agreements are available under standard academic terms.