AI photo apps have rapidly took over social media as the technology improved and hit the mainstream.
From people creating AI generated avatars of themselves to 90s inspired yearbook photos, professional headshots, and even photos of what future babies may look like, AI photo apps have reached peak popularity.
However, this fad is not altogether wholesome – AI photo apps, like many social media applications, track user data, and even use uploaded photos to train their AI model.
A new study, conducted by mobile phone retailer Fonehouse, has analysed twenty of the most popular AI photo apps to reveal the most intrusive to privacy and what happens to user data.
Using Apple’s app store, Fonehouse investigated the privacy labels for each app, which consists of 14 different categories, and investigated just how invasive these AI apps are
Training AI Tools
AI relies on data, be that image data, audio data, or text data. In order to better train AI models, developers often incorporate AI-generated material and uploaded data to better refine AI models. This is no different if a user inputs a sentence of uploads a phot.
Fonehouse found that over a third of apps say that they use user data to help train their AI tools or machine learning.
Only five percent clearly state they won’t do this, but 60% of apps were also unclear or did not specify what they do with user uploaded data.
Are They Keeping Your Selfies to Themselves?
Further, one in four AI photo apps store a user’s facial data after an image is created.
While 60% of apps delete facial data after an image is created and no longer store this data on their service, a quarter do, and 15% do not disclose what they do with the data or how long it is stored.
This data can be stored for different amounts of time depending on the app, ranging from no longer than seven days to up to three years from the data an account becomes inactive.
Alarmingly, the study also found that 75% of AI photo apps require users to grant a license within the terms and conditions to allow the app to use your content for a range of promotional purposes.
Some apps even state that by agreeing to the terms and conditions, users are granting them permission to use private photos in this way for perpetuity.
Recommended reading
- Face Off: AI Faces More Realistic Than Actual Humans, Study Finds
- The 10 Most Popular AI Tools of 2023
- Britain to Crack Down on Unauthorised AI Data Collection
What Apps Track the Most Data?
Based on how much data the apps collect overall, the app which came out on top was Photoleap, which collected eight out of the 14 categories (57%) of personal data.
When using the app, it collects contact info, user content, identifiers, diagnostics, usage data, purchases, location and other data categories.
This was followed by Picsart, Facetune, and Videoleap apps, which all collect data across seven categories.
Is Your Data Secure?
The study also used the Google Play App Store to analyse the security practices for each app too.
Some apps are designed to let users transfer their data to another site or service, mainly to share across social platforms like Instagram and TikTok.
These apps may declare in their data safety section that your data is transferred over a secure connection, however, one in five AI photo apps do not encrypt data that it collects or shares while its in transit.





