Fact check

Police use facial recognition to scan one million faces in WA trial leading to 122 arrests

Not enough data

The score shows how well the checked facts are backed by independent sources at the time of the check. It is not a verdict of absolute truth, and not a rating of the publisher.

Too few checkable facts for a score How is this calculated?

10 facts found in the article, 8 checked

1 confirmed
1 contradicted
6 could not be verified
2 not checked

Article published 2 days ago

Checked 2 days ago with claude-haiku-5-5 · Method 2026-10-10.

Contradicted by other sources

Contradicted The Western Australia police facial recognition trial generated nine incorrect alerts.

AI The Conversation reports that the WA police facial recognition trial produced eight incorrect alerts, not nine. Other results mention only two false alerts in the first week, which is a partial count and does not contradict the total.
- theconversation.com WA Police’s facial recognition trial delivered strong results. But was it an ethical success? contradicts

Verified facts

Confirmed Under the trial guidelines, cameras mounted on or near marked police vehicles scanned faces in designated public areas and matched them against an alert list drawn from police databases.

AI The WA Police official page describes a marked police van with live facial recognition cameras scanning faces at a chosen location and comparing them to a predetermined alert list, and news and policy sources report the same van-based setup. The claim's detail that the alert list was drawn from police databases is only loosely covered (the sources say 'watchlists' and 'alert list' without detailing the database origin), but the core of the claim is supported.

Could not verify

No independent source settled these either way. That doesn’t mean they are wrong.

Unverified WA police's five-month facial recognition trial, which concluded on a Saturday, scanned 1,380,882 faces across 100 deployments in 43 regional and metropolitan locations in Western Australia.

AI The results cover the trial only in its early weeks (about 131,000 faces scanned five weeks after the June 22, 2026 launch). None of them reports a five-month conclusion, the 1,380,882 total, 100 deployments, or 43 locations, so they neither support nor contradict the core figures.
Unverified The Western Australia facial recognition trial resulted in 122 arrests.

AI The results describe the Western Australia live facial recognition trial, reporting 33 alerts that sent officers toward a person and one false positive, but none gives an arrest count. None explicitly states 122 arrests or any other number of arrests, so the core figure can be neither confirmed nor contradicted.
Unverified The Western Australia facial recognition trial led officers to engage with 146 registered sex offenders.

AI None of the search results mention a Western Australia facial recognition trial or any engagement with registered sex offenders in that context. The only match for "146" is in an unrelated US law review article, so the core claim can't be checked, and silence is not a contradiction.
Unverified Three operations at the Perth Royal Show scanned 223,749 faces, resulting in 15 alerts, three arrests and seven interactions with registered sex offenders.

AI None of the search results address the Perth Royal Show facial recognition scanning; the only numeric overlap (223,749 in a Maine legislature document and a Wellington College financial table) is unrelated to the claim's core event, people, or figures.
Unverified WA police commissioner Col Blanch said on a Sunday media conference that officers engaged with each individual who was subject to a false alert, and that no one was arrested or made a complaint over the false alerts.

AI The snippets show Blanch describing the trial and stating that an algorithmic alert alone cannot justify an arrest, but none mentions a Sunday media conference or says officers engaged with each person subject to a false alert or that no one was arrested or complained over false alerts. The results do not address the core of the claim, and the snippets are truncated.
Unverified Civil Liberties Australia chief executive Bill Rowlings said WA police were "severely gilding the lily" by claiming the incorrect alert rate was 0.00065%.

AI None of the search results mention Bill Rowlings, Civil Liberties Australia, WA police, the 'gilding the lily' remark, or the 0.00065% incorrect alert rate. The only 'gilding of the lily' phrase appears in a UK press inquiry report about an unrelated book, so the results do not address the core of the claim.

Not checked

We check a limited number of facts per article to keep the service free. These don’t count towards the score.

The surveillance vans have been removed from public deployment while WA police conduct an internal evaluation of the trial.
The WA Office of the Information Commissioner said it was not involved in implementing the trial and warned of "collective privacy harms".

Author’s assumptions

Predictions or inferences in the article. They can’t be fact-checked.

The technology could disproportionately affect marginalised groups who are already overrepresented on police watchlists.
The trial has sparked significant concern about privacy and algorithmic bias.

Opinions

Views and judgements, the author’s or someone quoted. Not fact-checked.

The facial recognition trial in Western Australia is a serious threat to civil liberties.
The police's error rate framing is misleading and the real incorrect rate will only be known after court outcomes.

How we handle corrections

Report an error

What is it about?

What’s wrong? The verdict is wrong A source is wrong or unreliable Important context is missing Something else

Details (optional)