privacydeep-dive

AI Personal Trainer Apps Want Your Progress Photos

AI fitness coaching apps request body photos for personalization, then store them on company servers. Here's what the privacy policies actually say.

Body photos are some of the most sensitive images a person can take. They capture your physical form at a specific time, often at your most vulnerable — before a transformation, in the middle of a health struggle, or when you’re working toward a goal you haven’t reached yet.

AI fitness coaching apps want exactly these photos. They ask for them upfront, as part of onboarding. They call them “progress photos” or “baseline photos” or “form check images.” And then, by default, they keep them.


Why AI Fitness Apps Request Your Body Photos

The pitch makes sense: an AI system that can see your current body composition can give you more personalized training recommendations than one that can’t. If a coach can see that you have lower back curvature or shoulder imbalance, they can adjust your program accordingly. If they can see where you’re starting from, they can track whether the program is working.

AI fitness coaching apps like Future, Caliber, Whoop’s coaching features, and various newer entrants in the AI personal trainer category have extended this logic to automated systems. Upload a few photos — front, back, side — and the AI can identify your body type, estimate muscle mass distribution, flag postural concerns, and tailor a program specifically for you.

The feature is genuinely useful. But useful features require data, and the data being requested here is not abstract. It’s photographs of your body.


Where Those Photos Actually Go

When you take a progress photo in a fitness app and submit it, it leaves your device and travels to the app’s servers. This is unremarkable from a technical standpoint — it’s how app uploads work. But most users think of uploading a photo to an app as similar to showing it to someone, rather than as giving an indefinite copy to a company.

The destination matters. App server infrastructure is a very different environment from your phone’s local storage. Your phone’s photos are governed by your device’s own privacy settings, accessible to you, backed up to your personal cloud account. App servers are governed by the company’s infrastructure policies, retention schedules, security practices, and legal terms.

In most AI fitness app terms of service, photos uploaded to the app are licensed to the company for the purpose of providing the service. What “the service” encompasses is often defined broadly. Several apps reviewed in 2026 include clauses allowing uploaded photos to be used to improve or train AI models — sometimes under a separate opt-in, sometimes as a default that requires an active opt-out.

A progress photo submitted to help personalize your workout plan can, depending on the terms, become labeled training data for a body-composition AI model.


The Biometric Problem

Body photos are not definitionally biometric data under every legal definition, but they’re close. And in some jurisdictions, AI analysis of body photos for the purpose of estimating body composition, identifying physical characteristics, or tracking physical change over time is considered biometric processing.

Under the GDPR, biometric data processed for the purpose of uniquely identifying a natural person is a special category of data requiring explicit consent and additional protections. Illinois’ Biometric Information Privacy Act (BIPA), one of the most stringent US state laws, requires explicit informed consent before collecting “scans of face geometry” — and AI analysis of a face in a progress photo could qualify.

Several states have expanded biometric privacy laws in 2025-2026, and the question of whether AI body analysis constitutes biometric processing is actively litigated in some jurisdictions.

Most users uploading progress photos to fitness apps have no idea any of this applies to them. The consent they give is through a terms of service click-through that few read in full.


”Delete Your Account” Doesn’t Delete Your Photos

A consistent pattern in fitness app privacy policies: account deletion removes your account and stops your subscription, but doesn’t necessarily delete all the data the company holds about you.

Some apps distinguish between “account data” (which is deleted) and “aggregated or anonymized data” (which is retained). Progress photos that have been processed by an AI model may, depending on the app’s policy, exist in some form in model weights or training datasets even after the original images are deleted from your account.

More concretely: many apps state that deleted data is removed from active systems within 30 to 90 days, but may be retained in backup systems for longer. The exact retention schedule for photos specifically — as opposed to workout logs, which are less sensitive — varies and is often not disclosed.

If you’ve used an AI fitness app and uploaded body photos, it’s worth actually reading that app’s privacy policy section on data retention and deletion, not just clicking “delete account.”


The Specific Privacy Concerns with Progress Photos

Progress photos are a special category of personal image for several reasons.

They’re systematic and timestamped. A progress photo is designed to be part of a sequence — before and after, month by month. That sequence is more revealing than any individual photo because it shows change over time, which can reveal health conditions, medications, pregnancies, or injuries that no single photo would.

They’re often labeled. Many apps ask users to note weight, measurements, or health status alongside progress photos. This attaches highly personal health metadata to the image, creating a combination that is far more sensitive than either the photo or the measurements alone.

They’re taken in conditions of vulnerability. People take progress photos when they’re working on something — usually a body they feel uncomfortable with. The context in which they’re taken means they often capture users in ways those users would not want shared or stored.

They’re a target for data breaches. Health and fitness companies are attractive targets for data breaches because the data they hold is sensitive enough to be useful for extortion. Several fitness app breaches in recent years have included not just email addresses and payment data but workout histories and profile photos.


What Responsible Privacy Practices Would Look Like

Not all AI fitness apps handle this data the same way. The variance is significant.

Apps that handle this well typically:

  • Process body photos locally on-device rather than uploading to servers
  • Offer a meaningful, easy-to-find opt-out of any AI training use of photos
  • Provide specific retention timelines for photos separate from other account data
  • Distinguish between the data needed to provide the service and the data used to improve models
  • Are explicit in their privacy policies about whether photos are shared with third-party AI providers for processing

Apps that handle this poorly:

  • Upload photos to cloud infrastructure without specifying what happens to them
  • Include broad data use permissions in terms of service that cover AI training by default
  • Retain photos indefinitely without a clear deletion mechanism
  • Don’t distinguish between photos and less-sensitive data types in their privacy disclosures

The difference usually comes down to whether the company thinks of user privacy as a design constraint or as a compliance box to check.


What to Do Before Uploading

If you’re using or considering an AI fitness app that requests body photos, these are the questions worth answering before you upload:

Where are the photos stored? On-device processing is significantly more private than cloud upload. If the app requires cloud upload to function, ask why — many body analysis tasks are technically feasible on modern smartphone hardware without server-side processing.

Can the photos be used for AI training? Read the relevant section of the privacy policy, not just the summary. “Used to improve our services” is a common formulation that can mean AI training.

How do you delete them? Test whether there’s a way to delete uploaded photos independently of your account. If deletion is not granular — if you can only delete your entire account, not specific uploaded content — that’s a meaningful limitation on your ability to withdraw consent.

What happens after you cancel? Many fitness subscriptions are cancelled after a few months when users stop using the app. That’s precisely when users stop thinking about the data they left behind. Check whether cancellation triggers data deletion or whether your photos remain in the company’s systems indefinitely.


Storing Fitness Data Without Giving It Away

There’s a real reason to document your fitness progress. Tracking where you started and how you’ve changed is motivating, informative, and worth keeping. The question is where you keep it.

Progress photos stored locally on your device are accessible only to you (and any app with camera roll permissions). Progress photos stored in a private, encrypted storage service remain accessible to you across devices without being uploaded to a fitness company’s infrastructure.

daftei is built for exactly this kind of personal documentation: photos and files that matter to you, stored with AES-256 encryption at rest, transmitted with TLS 1.3, without being used to train any AI model — daftei’s or anyone else’s. Your progress photos are yours. They don’t become training data when you upload them, and there are no algorithms analyzing them for purposes other than letting you retrieve them.

That’s a different relationship than what most AI fitness apps offer. It’s the right relationship for data this sensitive.


The Larger Pattern: AI Features Funded by Personal Data

The AI fitness coaching market has grown substantially as general-purpose AI technology has become cheap enough to incorporate into consumer apps. The marginal cost of adding an AI feature has dropped. The incentive to collect data to improve that AI has not.

Progress photos are genuinely useful training data for body-composition AI models — they’re labeled (users tell the app their weight, measurements, and goals), they’re systematic, and they’re of high personal value to the users who submit them. From a machine learning perspective, a database of progress photos is a valuable asset.

The incentive for apps to collect this data, and the incentive to retain it broadly and use it for AI training, points in the same direction. Users who assume their progress photos are just used to customize their workout plan are often operating on an assumption the app’s terms of service doesn’t support.

Reading the privacy policy before uploading body photos is not paranoia. It’s the only way to know what you’re actually consenting to.

Your memories deserve better than an ad platform.

Try daftei free →
← All posts