Subscore

Data UseTesting methodology

Data Use looks at what the company does with your chats, photos, and personal information.

AI girlfriend apps can collect extremely private data. We check whether your information may be used to train AI, read by people, shared with other companies, or used for advertising.

We also check how long the company says it keeps your data and whether its privacy policy gives clear answers.

6 evidence groups
6 scored tests

Data Use is organized into 6 evidence groups. Each group contains one scored test: Training, Human Review, Data Sharing, Advertising, Retention, and Policy Clarity.

How the Data Use score is calculated

Every test gets a score from 0 to 10.

We multiply each test score by how much it counts. We then add all the points together to calculate the final Data Use score.

Data Use makes up 31% of the Privacy score.

The stored weights add up to 98, so we convert them into percentages that add up to 100%.

Policy Clarity has the biggest effect on this score because clear information is especially important when dealing with private data. Data Use makes up 31% of the Privacy score.

14.29%Training
14.29%Human Review
4.08%Data Sharing
4.08%Advertising
20.41%Retention
42.86%Policy Clarity

Weighted evidence groups (combined 100%)Data Use score31% of Privacy score

View exact calculation

Each test gets a score from 0 to 10. We multiply that score by how much the test counts. We then add all the points together.

Data Use makes up 31% of the Privacy score. Privacy makes up 10% of the overall performance score.

Scored tests and weights

Scored test	How much it counts	See scoring
Training	14.29%	View
Human Review	14.29%	View
Data Sharing	4.08%	View
Advertising	4.08%	View
Retention	20.41%	View
Policy Clarity	42.86%	View
Total	100%

How the score is calculated

Each test gets a score from 0–10
Test score × how much it counts
Example: 10 × 14.29% = 1.43 points
We do this for every test
We add all the points together
Final Data Use score
Counts for 31% of Privacy

Example calculation

We multiply each test score by how much it counts. We then add all the points together.

Scored test	Test score	How much it counts	Calculation	Points added
Training	10.00	14.29%	10.00 × 14.29%	1.43
Human Review	10.00	14.29%	10.00 × 14.29%	1.43
Data Sharing	4.00	4.08%	4.00 × 4.08%	0.16
Advertising	10.00	4.08%	10.00 × 4.08%	0.41
Retention	8.00	20.41%	8.00 × 20.41%	1.63
Policy Clarity	8.33	42.86%	8.33 × 42.86%	3.57
Final Data Use score	—	100%	Add all points	8.63/10

Special cases

Not Applicable

If a test does not apply, we remove it and spread its weight across the remaining tests.

Unknown

Privacy works differently from most other categories. If the company does not give a clear answer, we mark the result as Unknown and leave it out of the calculation. Unknown does not mean the app is safe. It means we could not confirm the answer.

Manual adjustment

In rare cases, we may adjust a score when the calculated result is clearly misleading. We always record the reason.

Why Data Use matters

AI girlfriend apps are adult platforms, and users may share very private information with them.

This can include intimate chats, personal photos, relationship details, and information you would never want shared publicly.

The problem is that it is not always clear what happens after you press Send.

Some companies may use chats or photos to train their AI. Employees or outside contractors may be allowed to review conversations. Personal data may also be shared with other companies or used for advertising.

That does not automatically mean the app is doing something illegal. But you deserve to know what is happening before you share anything private.

A strong Data Use result means the company clearly explains what it does, avoids unnecessary use of private information, and does not hide important details inside vague legal language.

What good Data Use looks like

A high Data Use score means the company clearly explains how it handles your information and avoids using it in ways that could put your privacy at risk. A lower score means chats or photos may be used for training, read by people, shared widely, or kept for an unclear amount of time.

Main benchmarks

GoodCautionWeak

AI training

NoUnknown or limitedYes

Human review

NoUnknown or limitedYes

Policy Clarity

80%+60–79%Under 60%

Minimum useful standard

Training: No
Human Review: No
Data Sharing: No
Advertising: No
Retention: Storage periods are given for important types of data
Policy Clarity: 80%+

Common red flags

The company says it may use chats for almost any business purpose
Employees or contractors may read conversations without clear limits
Data is shared with unnamed or very broad groups of companies
The policy does not say how long chats or deleted data are kept
Important answers are hidden behind vague wording
Different pages give conflicting privacy information

How we test Data Use

We use a paid test account and review the platform’s privacy policy, terms of service, help pages, account settings, and training and privacy controls.

We read the policy pages once and use the same sources for the related tests.

For each answer, we save the result, the source, the date checked, and any important wording or limitation.

We never assume a privacy protection exists just because it is common on other apps. The company needs to clearly explain it.

Training is also checked inside the test account because some apps place the real choice inside the settings instead of the privacy policy.

Limitations and edge cases

We cannot look inside a company’s private systems.

We can only check what the company says, what its settings show, and what normal users can test.

A company may follow better practices than its policy explains. It may also write a strong policy without following it properly. Our score reflects the evidence we can confirm.

Privacy wording can also be difficult to understand. We use the clearest reasonable meaning, save the source, and explain important uncertainty.

Policies change regularly. Our results show what the company stated on the recorded test date.

Evidence groups

Data Use has 6 evidence groups made up of 6 scored tests.

Training measures whether the company says your chats, photos, or other personal data may be used to train or improve its AI.

Many users may not expect private conversations to become training material.

Training

14.29%of Training (14.29% total)

Whether the company says your chats, photos, or other personal data may be used to train or improve its AI.

How we test

We search the privacy policy, terms of service, help pages, and account settings. We look for clear wording about AI training, model improvement, product improvement, and similar uses.

What counts

Chats used to train AI models
Photos used to improve image models
Personal content used to improve automated systems
Data used for training unless the user opts out
Clear wording that private content is not used for training

What does not count

Anonymous technical data used to fix crashes
Basic usage statistics with no chat content
Claims made by unofficial users or social-media accounts
Assumptions based on what other companies do

Result shown

No — the company says private chats and photos are not used for AI training

Source checked: Privacy policy

Scoring

We use the result categories below.

Result	Score
No — private content is not used for training	10/10
Limited — only some data is used or clear limits apply	3/10
Yes — private content may be used for training	0/10
Unknown — no clear answer	Excluded

In this example, Training scores 10/10.

Evidence group 2 of 6

Human Review measures whether employees or outside workers may read your chats.

Human access may sometimes be needed for support, safety, or abuse reports. The important part is whether the company clearly explains when and why it can happen.

Human Review

14.29%of Human Review (14.29% total)

Whether employees or outside workers may read your chats.

How we test

We search the privacy policy, terms, and help pages for wording about employees reading chats, contractors reviewing content, safety reviews, support access, quality checks, and moderation.

What counts

Employees may review conversations
Contractors may access chat content
Chats may be checked for safety or quality
Human access is clearly limited to specific situations
The company clearly states that people do not routinely read chats

What does not count

Automated moderation with no human access
A support agent reading a message you directly send to support
Unverified claims from users
General wording about “processing” data with no mention of people

Result shown

No — the company says chats are not routinely reviewed by people

Scoring

We use the result categories below.

Result	Score
No — people do not routinely review chats	10/10
Limited — review only happens in clearly limited situations	3/10
Yes — people may broadly review chats	0/10
Unknown — no clear answer	Excluded

In this example, Human Review scores 10/10.

Evidence group 3 of 6

Data Sharing measures whether personal information is shared with other companies.

Some sharing is needed to run an online service. The bigger concern is broad or poorly explained sharing.

Data Sharing

4.08%of Data Sharing (4.08% total)

Whether personal information is shared with other companies.

How we test

We review the list of third parties or groups of companies that may receive user data. We record whether sharing happens, why the data is shared, which types of companies receive it, and how many third-party categories are listed.

What counts

Cloud and hosting providers
Analytics companies
Payment processors
Advertising partners
AI model providers
Other companies receiving user information

What does not count

Data kept only inside the company
Sharing required by law when clearly explained
Anonymous statistics that cannot reasonably identify a user
Companies mentioned without receiving user data

Result shown

Limited — data is shared with hosting, analytics, and payment providers

Third-party categories: 3

Scoring

We use the result categories below.

Result	Score
No — personal data is not shared with outside companies	10/10
Limited — sharing is clearly restricted to necessary providers	4/10
Yes — personal data is shared more broadly	0/10
Unknown — no clear answer	Excluded

In this example, Data Sharing scores 4/10.

Evidence group 4 of 6

Advertising measures whether personal information is used for ads, personalized marketing, or user profiling.

This matters because private chats and interests should not quietly become advertising data.

Advertising

4.08%of Advertising (4.08% total)

Whether personal information is used for ads, personalized marketing, or user profiling.

How we test

We check the privacy policy and account settings for wording about advertising, personalized ads, marketing profiles, tracking, behavioral advertising, and selling or sharing data for ads.

What counts

Personal information used to choose ads
User behavior used for ad targeting
Data shared with advertising partners
Profiles created for marketing
A clear statement that personal data is not used for advertising

What does not count

The company sending its own basic service emails
A general newsletter the user chose to receive
Non-personalized ads with no user profiling
Necessary payment or account messages

Result shown

No — the company says personal data is not used for personalized advertising

Scoring

We use the result categories below.

Result	Score
No — personal data is not used for advertising	10/10
Limited — some restricted advertising use applies	3/10
Yes — personal data may be used for advertising	0/10
Unknown — no clear answer	Excluded

In this example, Advertising scores 10/10.

Evidence group 5 of 6

Retention measures how long the company says it keeps your information.

Deleting a chat or account does not always mean the data disappears immediately.

Retention

20.41%of Retention (20.41% total)

How long the company says it keeps your information.

How we test

We record the stated storage period for chats, account information, payment information, and deleted data. We use the exact time given by the company, such as days, months, or years.

What counts

A clear number of days, months, or years
Different periods for different types of data
A clear explanation of what happens after deletion
Legal or security exceptions that are clearly explained

What does not count

We keep data as long as needed with no useful limit
A deletion button with no explanation of what happens afterward
Guessing how long a company probably stores information
Retention periods from an outdated policy

Result shown

Chats: Deleted within 30 days

Account information: Kept while the account is active · Payment records: Up to 7 years · Deleted data backups: Up to 90 days

Scoring

Example Retention score: 8/10.

Retention receives a manual score from 0 to 10. Clear, specific, and reasonable storage periods lead to a higher score. Vague wording, missing periods, or very long unexplained storage lead to a lower score. The product evidence shows the recorded periods and the reason for the score.

Evidence group 6 of 6

Policy Clarity measures whether the company gives clear answers to the privacy questions users are most likely to care about.

A long privacy policy is not helpful when it avoids the important questions.

Policy Clarity

42.86%of Policy Clarity (42.86% total)

Whether the company gives clear answers to the privacy questions users are most likely to care about.

How we test

We check whether the company clearly answers these six questions: Are chats used for AI training? Can people read chats? Is data shared with other companies? Can users delete their data? How long is data stored? What security protection is used? Each clear answer counts as one passed check.

What counts

A direct Yes or No answer
A clear explanation of when something happens
Important restrictions stated in plain language
Information that is easy to find in official pages

What does not count

Vague statements such as we care about privacy
Legal wording that never answers the question
Important information spread across conflicting pages
Answers from unofficial sources
A topic being mentioned without explaining what actually happens

Result shown

5 of 6 questions clearly answered

Policy Clarity result: 83.3%

Scoring

A higher percentage means a higher score.

Questions clearly answered	Score
0 of 6	0/10
1 of 6	1.67/10
2 of 6	3.33/10
3 of 6	5/10
4 of 6	6.67/10
5 of 6	8.33/10
6 of 6	10/10

A result of 5 clear answers scores 8.33/10.

Back to Privacy