AI Girlfriend Coach
Home / Data
Data and downloads

The Data Behind the Research

Every file this site's research produces, in one place: the reports, the datasets, the chart artwork, and the headline tables in plain HTML so you can read the numbers without downloading anything.

If you want to know what the research found, Research is the page for that. This one is the file room.

Files

Downloads

Free to use with attribution. Nothing here asks for an email address.

FileFormatLicence
Master Research ReportThe full survey study: methodology, questionnaire, every table, subgroup analysis, statistical testing, data-quality screening and appendices. Version 1.0. PDF
41 pages
0.8 MB
Free with
attribution
Download
Visual Research DeckThe same study as a presentation: the major findings, the comparisons and the charts, ready to show. PPTX
28 slides
6.2 MB
Free with
attribution
Download
Charts and InfographicsAll seven survey charts as vector SVG and 2x PNG, each with the citation baked into the file so the credit travels with the chart. ZIP
15 files
0.6 MB
Free with
attribution
Download
AI Companion Funding Dataset30 companies and $415.6M in disclosed funding, with rounds, investors, dates and the source behind each entry. PDF
5 pages
16 KB
CC BY 4.0 Download
Need a cut that is not here? A companion workbook with every cross-tabulation and statistical test, plus an anonymized respondent-level dataset with the survey platform's metadata removed, is available on request for verification. Email ranji@aigirlfriend.coach.
The numbers

Survey Headline Tables

The five results most often asked for, in plain HTML. Every figure uses the 1,709 quality-screened responses unless a smaller base is named. Full tables, confidence intervals and significance tests are in the Master Research Report.

Use of AI girlfriend and boyfriend apps, and future interestQ: Have you ever used an AI girlfriend or AI boyfriend app? Would you consider using one in the future?
GroupRespondentsShare of sampleWould consider future use
Current users21112.3%96.2%
Former users21112.3%74.4%
Never used one1,28775.3%11.7%
Main reason for using an AI companionSingle choice, asked of all 1,709 respondents as a hypothetical rather than only of users. Sums to 100%.
ReasonShare
Someone to talk to, or feeling less alone56.6%
Improving mental health or coping with depression18.5%
Confidence or self-esteem14.6%
Affection, romance or sex10.3%
Would a partner's romantic or sexual relationship with an AI be cheating?Single choice, base 1,709. Sums to 100%. The first two rows together are the 50.5% headline figure.
AnswerShare
Yes, any romantic or sexual AI interaction counts33.4%
Yes, but only once emotional attachment develops17.1%
Not cheating, but I would be uncomfortable with it21.2%
Not cheating, an AI is not a real person28.3%
Would consider an AI girlfriend or boyfriend, by generationRespondents aged 81 and over (37) are excluded, below the report's 100-respondent reporting threshold. Separate bases, so these do not sum.
GenerationWould consider future use
Millennials44.7%
Gen X31.6%
Gen Z12.5%
Boomers8.6%
If single and looking for companionship, which would you rather use?Single choice, base 1,709. Sums to 100%.
AnswerShare
Dating apps52.2%
Unsure26.8%
Both13.2%
An AI girlfriend or boyfriend7.8%
One number worth reading twice. Excluding 441 low-quality responses cut reported current use from 30.3% to 12.3%, and the share saying they could genuinely fall in love with an AI from 27.1% to 8.3%. The screening rule and the reasoning are in section 3 of the report.
How to read it

Methodology and Limitations

The survey went to 2,150 U.S. adults aged 18 and over through SurveyMonkey Audience on 26 to 27 August 2026, with ten questions covering AI companion use, motivations, romantic attachment, cheating, and attitudes toward AI against human relationships. After quality screening, 1,709 responses were analyzed. The sample was balanced on age and gender against U.S. Census benchmarks, but not on region or household income.

Before quoting any figure here, three limits are worth knowing:

  • These are associations, not causes. The survey cannot show that AI companions cause or relieve loneliness, or that using one changes what a person believes about it.
  • Adoption figures are not national prevalence. Reported use differs sharply by device type in this sample, and the sample is not balanced on region or income, so the usage numbers describe these respondents rather than U.S. adults as a whole.
  • Small subgroups are not lead findings. Anything under 100 respondents is reported as context, never as a headline, and section 14 of the report flags which findings are device-sensitive.

The testing methodology covers the product side: how platforms are chosen, what gets tested, and how the paid rounds work.

Artwork

Charts and Reuse

All seven survey charts are in the ZIP above as vector SVG and 2x PNG. Each file carries its own source line, so a chart that travels arrives attributed. Use them in articles, decks and reports with the citation below. Please do not alter the figures or crop the source line off.

How to cite: AI Girlfriend Coach (2026). AI Romance and Modern Relationships: A Survey of 2,150 U.S. Adults. Master Research Report v1.0, August 2026. Data collected via SurveyMonkey Audience, 26 to 27 August 2026.

The funding dataset is separate and more permissive: it is published under CC BY 4.0, so it can be redistributed and built on commercially, with attribution.

Elsewhere

The Rest of the Evidence

Not everything is a file. The four standing studies, the 18 compiled statistics pages and the methods behind each of them are indexed on Research. The hands-on platform testing that all of it draws on, including the screenshots and the prices, sits in the living reviews, and the shortlist of AI girlfriend apps I actually recommend is on the homepage.