Why Face Data Is Different
Face data is biometric: it comes from your body. That makes it different from most personal data in three ways:
- You can't change it. A leaked password can be reset. A leaked face template can't.
- It's collected at a distance. Fingerprints need your finger. Your face can be captured by any camera, without you knowing.
- It links everything. A face template can connect a photo on a dating app, a CCTV frame and a social media post to the same person.
Put those together and a large face database becomes a tool for tracking people across their lives, which is why regulators treat it so seriously.
The Cases That Shaped the Debate
- Clearview AI. The company scraped billions of photos from social media and the web to build a face search tool sold to police. Regulators in several countries ruled it illegal under their privacy laws. In 2024 the Dutch data protection authority fined it €30.5 million.
- Facebook. In 2021, Facebook agreed to a $650 million settlement under Illinois' biometric privacy law over its photo-tagging feature, and later that year shut down its face recognition system and deleted more than a billion face templates.
- Police use. Several US cities, starting with San Francisco in 2019, banned police use of face recognition. Wrongful arrests linked to face searches, covered in bias in face recognition, fuelled that.
Read more: Bias in Face Recognition: What It Is and What to Do About It
The Laws That Protect You
| Law | Where | What it does |
|---|---|---|
| GDPR | EU, UK (as UK GDPR) | Treats biometric data used to identify people as a special category. Needs explicit consent or another narrow legal basis |
| BIPA (2008) | Illinois | Requires written consent before collecting face data, and lets individuals sue |
| CUBI | Texas | Requires consent and limits sale of biometric identifiers |
| CCPA / CPRA | California | Treats biometrics as sensitive personal information with rights to know and delete |
| EU AI Act (2024) | EU | Bans untargeted scraping of faces to build databases and restricts real-time face recognition in public |
Laws differ a lot by country and state, but the trend is clear: more consent, more limits, and more ways for people to say no.
What to Check Before Using a Face App
- Is the photo stored? For how long, and where?
- Is a face template kept? A stored template can be matched against other photos later.
- Is it used for training? Some apps use uploads to improve their models.
- Is it shared or sold? Look for third parties in the privacy policy.
- Can you delete it? And does the app say how?
- Can it identify strangers? Apps that let anyone upload any face and get a name are a red flag.
Read more: Face Recognition Accuracy Across Demographics: What the Research Shows
Get new guides by email
How Ollie Handles Your Photo
Here's exactly what happens, so you can check it against the list above:
- Your photo is processed in memory on the model server and never written to disk or logged. Once your results are returned, it's gone.
- Your face template (the 512 numbers) exists only for that search and isn't saved.
- Training: uploads are never used to train the model. It was trained on a public research dataset before launch.
- What is stored: for each search, an account ID (or a hashed IP address for guests), the IP address and a timestamp, to enforce free search limits. These are deleted after 90 days.
- No identification: Ollie matches you against a fixed list of public figures. It can't tell anyone who a stranger in a photo is.
The full details are in Ollie's privacy policy. Data that's never stored can't leak, which is the simplest privacy protection there is.
Your Rights Over Face Data
Depending on where you live, you may have the right to ask a company what face data it holds about you, to have it deleted, and to refuse its collection in the first place. Under GDPR, these are formal rights with deadlines for a response. If an app or service won't say what it does with your face, that alone is a reason not to use it.