Research - Super Recognisers better at detecting AI generated faces

/

The academic study notes, "The AI revolution has produced synthetic faces that often appear more human than real photos of people. We tested whether individual differences in face recognition abilities explain variation in discriminating AI from human faces. Super-recognisers—people with exceptional ability to recognise human faces (N = 36)—outperformed a standard participant group by 15 percent..."

Criminals and terrorists will utilise AI faces to commit fraud, breach international borders and further their aims. Using human Super Recogniser skills gives law enforcement and counter-terrorism units opportunities to defeat these tactics.

Super-Recognisers can Detect AI-hyperrealism | Sciety Labs (Experimental)

German Super Recogniser Captures Thief

/

This month, a German police Super Recogniser spotted a thief whose image had been captured by CCTV in the main railway station. The pickpocket, who had stolen an I-phone in December 2024, thought he had escaped detection, but his face had been remembered by an officer deployed on the S-Bahn.

BPOL-HH: Super recogniser of the Federal Police recognizes suspect after a crime in the ... | Press portal

US Super Recognizer

/

Metro Transit Police Officer Kache Fields is used as a Super Recognizer (as it is spelt in the States!) and has identified 1000 suspects. The Association is pleased to see SR skills being used in the USA. Full details in the attached article:

https://wjla.com/news/local/metro-transit-mtpd-dc-washington-phtographic-memory-fields-investigative-units-agencies-be-on-the-look-out-stations-rail-bus-investigation-arrests-weapons-safety

Make-Up Classes to Defeat Automated Facial Recognition

/

The Times of London newspaper has reported that activist make-up artists at the Glastonbury music festival have held a workshop to teach how to cause a glitch in facial recognition software, using a tactic known as “face breaking” or “computer vision dazzle”. The concept is based on dazzle camouflage used by ships during WW1, which used high-contrast colours and sloping lines to distort the structure of a vessel.

Combining human Super Recognisers and automated FR gives the best way of ensuring criminals and terrorists cannot utilise this tactic to escape justice. Do not over rely on technology - use the skills of your people too.

https://www.thetimes.com/culture/music/article/glastonbury-festival-greenpeace-how-to-beat-facial-recognition-protests-0v2c8cjl6

Super Recogniser Makes 3000 Identifications!

/

Excellent work by Police Community Support Officer (PCSO) Andy Pope of the West Midland Police in UK on making 3000 identifications from CCTV and other images. Andy was made an Honorary Fellow of the Association by the chairman, Lord Lingfield, in recognition of his work. BUT only 40% of his identifications result in suspects being charged, as the West Midlands Police, like many other law enforcement agencies, do not manage image identifications in the systematic manner that we see with other forensic disciplines.  Too many suspects are escaping justice because Andy’s top work is top being used effectively by senior police officers.

https://www.dailymail.co.uk/news/article-14816527/Super-recogniser-PCSO-Memory-Cop-breaks-record.html

Scottish Police Encouraged to Use Super Recognisers

/

In the article below, Prof Josh Davis of Greenwich University, as honorary fellow of the Associate encourages Police Scotland to use their human Super Recognisers.

From the article: Professor Davis said: “When you want to try and identify the perpetrator and you have a database of potential mugshots, that face is fed into a system and it searches through, potentially, a very large database if you’re going to put it through the entire police record of everyone who’s been arrested.  It brings up a large array of face potential matches and against each one you’ll get some sort of percentage likelihood that it’s a match.  But if that person hasn’t been arrested for five to ten years their appearance will have changed. AI is not as good at that point at all.  It’s AI and super-recognisers that work the best together, rather than one or the other.”

Inside little-known super-spotters who help catch criminals - just 2% of UK have gift, find out if you're one of them | The Scottish Sun

New Fictional Novel Features Super Recogniser

/

A new book “Her Last Move” by crime writer, John Marrs, features a Met Police Super Recogniser. From Amazon:

“Whatever you do, don’t read this in the dark…” —Cara Hunter, author of Close to Home

He hides in the shadows, waiting for the perfect moment. Each kill is calculated, planned and executed like clockwork.

Struggling to balance her personal and professional life, young DS Becca Vincent has landed the biggest case of her career – and she knows that it will make or break her. But how can she identify one face in a sea of thousands? With the help of Police Super Recogniser Joe Russell, she strives to catch a glimpse of the elusive murderer, but he’s watching her every move.

Time is not on their side. The body count is rising, and the attacks are striking closer and closer to home. Can Becca and Joe uncover the connection between the murders before the killer strikes the last name from his list?

From the bestselling author of When You Disappeared and The Good Samaritan comes his most thrilling novel yet.

Her Last Move eBook : Marrs, John: Amazon.co.uk: Kindle Store

Ireland & Police Use of Facial Recognition

/

The police in Ireland will be using live and retrospective computerised facial recognition, but who will verify these identifications? In the UK, this is covered by the Surveillance Camera Commissioner Code of Practice (Amended Nov 2021).

The key part is Para 2.4, which notes that, “Any use of facial recognition or other biometric characteristic recognition systems needs to be clearly justified and proportionate in meeting the stated purpose and be suitably validated. It should always involve human intervention before decisions are taken that affect an individual adversely.”    

The best people to verify computer identifications are human Super Recognisers. Combining the best humans with the computer provides excellent results.

https://www.irishtimes.com/politics/2025/04/25/work-on-legislation-to-introduce-facial-recognition-technology-well-advanced-says-minister/