Clearview AI Loses BC Appeal: Face Scraping Rules [2026]
In February 2026 the British Columbia Court of Appeal rejected Clearview AI's bid to escape Canadian privacy law, upholding orders to stop collecting and to delete images of B.C. residents. This guide explains the ruling, why scraping public photos is a privacy problem, and how to stop your faces from ending up in a database.

Clearview AI built a face search engine by collecting photos from across the public web, reportedly more than 30 billion of them. In February 2026 a Canadian appeals court refused to let the company off the hook. The British Columbia Court of Appeal ruled that Clearview must comply with provincial privacy law even though it stopped offering services in Canada, and that orders to stop collecting and to delete images of B.C. residents stand.
This guide explains what the court decided and why, what it means for the photos and videos you post, and the practical steps that actually reduce your exposure to face scraping. If you publish video with other people's faces in it, BGBlur's face blur covers them before scrapers can.
TL;DR: The Clearview AI B.C. Ruling
| Question | Answer |
|---|---|
| Which court? | British Columbia Court of Appeal |
| When? | February 20, 2026 |
| Who wrote it? | Justice Karen Horsman, for Justices Nitya Iyer and Paul Riley |
| Result? | Clearview's appeal dismissed; 2024 B.C. Supreme Court decision upheld |
| Which law? | British Columbia's Personal Information Protection Act (PIPA) |
| What was ordered? | Stop collecting, using and disclosing B.C. residents' images and delete previously scraped ones |
| How big is the database? | Over 30 billion faces, per Biometric Update |
| What can I do? | Blur faces before you post; the ruling can't retrieve images already collected |
What Did the B.C. Court of Appeal Decide?
According to Biometric Update's report on the decision, the court dismissed Clearview's appeal and upheld a 2024 B.C. Supreme Court ruling that the company violated the province's Personal Information Protection Act. That ruling in turn confirmed orders issued in 2021 by B.C.'s privacy commissioner, who found that Clearview collected images of residents without consent. The orders required the company to:
- Stop collecting, using and disclosing images of British Columbians
- Delete images of B.C. residents it had already scraped
The case started in 2020 when the privacy commissioners of British Columbia, Alberta, Québec and Canada jointly investigated Clearview after it began marketing its services to Canadian customers. Clearview left the Canadian market during the probe, according to national coverage of the appeal.
Why Did Clearview Argue It Didn't Have to Comply?
Clearview made four arguments, all of which the court rejected:
- No provincial connection. It said PIPA didn't apply because it had no real link to British Columbia.
- Public means free to take. It said posts on public social media don't require consent to collect.
- Can't police the global web. It said it couldn't be forced to shield one province from a worldwide scraping operation.
- Can't un-collect. It said it couldn't retroactively remove images already gathered.
Justice Horsman's answer to the first was the one that will echo. She wrote that accepting Clearview's position would mean "any company that acquires personal information on the internet using a global search engine" would be immune from domestic privacy laws, significantly compromising a jurisdiction's ability to protect personal information online. The court also noted that Clearview had complied with an Illinois biometric-privacy ruling, which suggested regional compliance is technically possible.

Why Is Scraping Public Photos a Privacy Problem?
Because a face is not like other data. You can change a password, but you can't change your face, and once a faceprint sits in a searchable database it can be used to identify you from a single photo taken in the street. Clearview's customers, including law-enforcement agencies, can upload an image and look for matches across billions of scraped pictures.
"It was public" is the argument regulators keep rejecting. A photo posted for friends is not consent to have your face templated and sold. That view is reflected in the EU's rules, where the AI Act bans building facial recognition databases through untargeted scraping, covered in our guide to the EU AI Act facial recognition scraping ban. Canada's federal approach is covered in our PIPEDA video privacy guide.
Does the Ruling Get My Photos Back?
No, and this is the honest part. The decision confirms that a Canadian regulator can order deletion of images of B.C. residents. It does not automatically delete your face from every copy of every database, and it does not bind courts outside British Columbia. Enforcement against a foreign company that no longer operates in the market has always been the hard part, which is why the court's reasoning about jurisdiction matters so much.
For everyone else, the legal outcome is slower than the technology. The photo you posted last year may already have been collected. That leaves prevention as the only option that works today.
How Do I Stop My Faces From Being Scraped?
You can't stop a scraper from visiting a public page, but you can control what it finds. Layer these steps:
- Blur faces before you post. This is the one that removes the usable face from that file. See how to blur faces in video with AI.
- Blur other people first. Bystanders, customers, students and children never agreed to appear online. Our post on whether someone can record and post your face explains the consent rules.
- Tighten account privacy. Private accounts and friends-only posts are not scraped from the public web in the same way.
- Strip metadata. Location and device data in file metadata can identify where a photo was taken.
- Don't trust cheap disguises. Patterned clothing and makeup are not reliable, as we cover in digital camouflage and face detection.
How to blur faces before posting
- Open BGBlur and upload your photo or video (MP4, MOV, M4V or common image formats).
- Let the AI detect every face. Motion tracking keeps moving people covered across the whole clip.
- Preview the result and check crowds and background figures.
- Export. Processed files are deleted within 24 hours. For footage you'd rather not upload, the BGBlur desktop app blurs faces on your own computer.
How Did the Clearview Case Get Here?
The February 2026 decision is the latest step in a case that has run for years. Based on the reporting by Biometric Update and CP24, the sequence looks like this:
- 2020: The privacy commissioners of British Columbia, Alberta, Québec and Canada open an investigation after Clearview begins marketing its facial recognition service to Canadian clients.
- During the investigation: Clearview withdraws from the Canadian market.
- 2021: B.C.'s privacy commissioner orders the company to stop offering its services in the province, stop collecting, using and disclosing residents' images, and delete images already scraped.
- 2024: The B.C. Supreme Court upholds the finding that Clearview violated the Personal Information Protection Act.
- February 20, 2026: The B.C. Court of Appeal dismisses Clearview's appeal.
Notice what didn't change across those five years: the underlying image collection had already happened. Every legal step confirmed that it was unlawful without physically retrieving the data. That gap between a legal finding and a technical remedy is the main reason prevention matters more than enforcement for ordinary people.
What Does the Ruling Mean for Businesses That Publish Video?
If your company posts footage of customers, students, patients, employees or event attendees, the ruling reinforces three points.
- Public posting isn't a consent shortcut. Clearview argued that public social media posts don't require consent to collect. The court didn't accept that framing, and regulators in other jurisdictions have taken a similar view of scraping.
- Where you are doesn't fully protect you from where they are. The court's reasoning was that a company can't evade a jurisdiction's privacy law just because it operates from elsewhere or has left the market. For a business, this cuts both ways: your obligations to people in a region can follow your footage there.
- Your published footage is raw material. Anything you publish with identifiable faces can end up in a scraper's index, so the safest publishing policy is to treat unblurred faces as a permanent disclosure.
A workable policy is short: default to blurring faces of anyone who hasn't agreed to appear, keep the originals in a controlled location, and publish only the processed version. For sensitive footage, do the blur on your own hardware, as described in our guide to blurring faces in video without uploading it.
Who Should Care Most?
Parents and schools: Children's photos posted publicly are a common source of scraped faces. See how to blur kids' faces before posting.
Journalists and activists: Faces at protests can be matched to identities, a risk covered in our protest privacy guide.
Businesses: Customer and employee footage published on social media creates biometric exposure for the people in it and compliance risk for you.
Creators: Vlogs and street content capture strangers by default.
Honest Limitations
Blurring protects the files you publish. It doesn't undo images already collected, and weak pixelation can be partly reversed, so use a strong blur and review the export. The B.C. ruling applies to British Columbia law and does not bind courts elsewhere. Also, Clearview is one of many face-search services, so a ruling against one company doesn't remove the technology. This article is general information, not legal advice.
The Bottom Line
The B.C. Court of Appeal told Clearview AI that leaving the market doesn't erase its obligations under Canadian privacy law, and that a global scraping operation can't shelter behind borders. It is a meaningful precedent for regulators, but it can't retrieve faces already collected. The step within your control is upstream: blur faces before you post. Try BGBlur's face blur on your next clip.