The successful Thames Valley Police use of human Super Recognisers to target suspects engaged in VAWG (Violence Against Women & Girls) and other offenders featured in yesterday's Media Day and have been reported on the BBC, as well as local media stations. For a report see this link:
https://www.bbc.com/news/uk-england-oxfordshire-66609161
Here is an academic view on human Super Recognisers from Professor Josh Davis of Greenwich University as featured on The Perception Podcast. Listen to Josh as he explains how he began his research and his latest findings.
https://omny.fm/shows/the-perception-podcast/professor-josh-davis
From the article:
Porcha Woodruff was eight months pregnant when police in Detroit, Michigan came to arrest her on charges of carjacking and robbery. She was getting her two children ready for school when six police officers knocked on her door and presented her with an arrest warrant. She thought it was a prank.
“Are you kidding, carjacking? Do you see that I am eight months pregnant?” the lawsuit Woodruff filed against Detroit police reads. She sent her children upstairs to tell her fiance that “Mommy’s going to jail”.
…Woodruff later found out that she was the latest victim of false identification by facial recognition. After her image was incorrectly matched to video footage of a woman at the gas station where the carjacking took place, her picture was shown to the victim in a photo lineup. According to the lawsuit, the victim allegedly chose Woodruff’s picture as the woman who was associated with the perpetrator of the robbery. Nowhere in the investigator’s report did it say the woman in the video footage was pregnant. A month later the charges were dismissed due to insufficient evidence (full article link is below).
The view from the Association: In the UK and EU, no arrests can be made from an automated facial recognition (AFR) identification unless there has been "human intervention". The best person to verify an identification is a Super Recogniser - the US agency NIST stated that combining AFR with a Super Recogniser produced "near perfect results". If police agencies have invested in computer facial recognition, they should ensure that they also invest in the best humans to make it even more effective. Reports such as those quoted will result in more restrictions and even bans on the use of AI by security and law enforcement...so this Association strongly recommends the use of the best people to check identifications before making arrests.
The CEO of the Association contacted all English Premier League football clubs as follows:
Premiership matches (and other sporting events) face three safety challenges this season and I believe that my team can assist you. These challenges are:
1. Preventing “Just Stop Oil” from fulfilling their threat to disrupt at least one match per week.
2. Demonstrating efforts to comply with the forthcoming Martyn’s Law (Protect Duty).
3. Ensuring banned fans do not enter and engage in illegal chanting, racist comments or other criminal behaviour.
SRI Ltd are experts in selecting, training and providing human Super Recognisers. These are people with a natural (and exceptional) ability to spot faces. I can establish which of your stewards (or other staff) have this exceptional skill (assessed by academic experts) and train them to use it in an operational environment, complying with the law. These Super Recognisers can be then deployed to critical entry points, the CCTV control room or other key positions to spot those banned or intent on causing crime or disruption. I can also provide experienced Super Recognisers to assist them. If you are using computer facial recognition, then Super Recognisers will enhance this technology, if not, they are a far cheaper alternative.
Details of the threat from JSO are in this LBC story:
'The Prem's our Holy Grail': Just Stop Oil plan to disrupt a Premier League game every... - LBC
Identification parade evidence, especially from members of the public, who have just been the victim or witness of a dreadful crime (a very stressful event), is fraught with danger. Such is the problem, that judges are required to give a warning to juries (in UK based on R v Turnbull). The case below highlights how the wrong person can be sent to jail on such evidence. Should we not test the ability of a witness as to their skill at making identifications? This could show the value to place on their evidence - allowing investigators and prosecutors to make informed decisions. Our Super Recogniser tests identify not only those who are very good at recognising faces, but also those with average or very poor ability. This could assist law enforcement AND defence counsel. Justice can benefit from this.
https://news.sky.com/story/andrew-malkinson-man-wrongly-jailed-for-17-years-hits-out-at-lying-police-and-says-he-was-kidnapped-by-the-state-12928036
Click here for further details
Thames Valley Police are leading the way in using Super Recognisers in the fight against the completely unacceptable levels of violence against women and girls. The UK Home Secretary has made tackling VAWG a key policing priority. The report below featured in Policing Insight. The text is pasted below.
Sexual harassment and sexual assault are worryingly common features of the night-time economy for many women and girls; now a novel policing tactic being utilised by Thames Valley Police, codenamed Project Vigilant, is seeking to address the problem with a preventative approach that exploits the unique talents of ‘super recognisers’, as Policing Insight’s James Sweetland reports.
It’s a depressing fact that the so-called night-time economy, whether pubs, bars, clubs or other spaces, is the setting for constant sexual harassment – as well as even more serious sexual offences.
One tactic being deployed by Thames Valley Police suggests a novel way policing can address VAWG, with so-called ‘super recognisers’ playing an important role.
This is, of course, no new problem. A 2018 report from the House of Commons’ Women and Equalities Committee described women “being routinely harassed in bars and clubs at night to the extent that it is the norm on a night out”.
Adding to this, the Committee noted: “Our own research shows that sexual harassment is regarded as more socially acceptable when it takes place in a bar than in the street or in public generally, with one participant describing it as ‘part of the unwritten rules of a bar’.”
Recognising the need to address these issues, last week the Government added Violence Against Women and Girls (VAWG) to the strategic policing requirement, meaning police forces must treat these crimes as a top-tier, national threat.
Policing will now be under pressure to take tougher, more effective action, and must be more innovative to really make a difference. One interesting tactic being deployed by Thames Valley Police – codenamed Project Vigilant – suggests a novel way policing can address VAWG, with so-called ‘super recognisers’ playing an important role.
Super recognisers
Many in policing may have come across the term ‘super recognisers’ before. It was coined back in 2009 by academics at Harvard University and UCL’s Institute of Cognitive Neuroscience who ran tests on “four people who claimed to have significantly better than ordinary face recognition ability”. The results: “Exceptional ability… confirmed in each case.”
One super recogniser at West Midlands Police has spotted over 2,000 suspects since 2012, while others have been deployed at major events such as Notting Hill Carnival; they are even credited with helping catch the Russian agents responsible for the Salisbury attacks.
From this, the idea of ‘super recognisers’ was born, describing individuals with the opposite of prosopagnosia – “a condition marked by exceptionally poor face recognition ability”. Since then, another leading academic, Professor Josh Davis, has developed the field considerably.
Super recognisers – whether police officers or police staff – have been identified and deployed by many different forces. One super recogniser at West Midlands Police has spotted over 2,000 suspects since 2012, while others have been deployed at major events such as the Notting Hill Carnival; they are even credited with helping catch the Russian agents responsible for the Salisbury attacks.
With a growing debate around the role of artificial intelligence (AI) to recognise faces through live facial recognition, super recognisers offer an interesting human alternative. In an article for Policing Insight in 2018, Mike Neville, the CEO of Super Recognisers International, addressed exactly this.
He argued that: “Super recognition is a force multiplier adding value to CCTV capability. Super recognisers are not in competition with facial recognition systems, but work in parallel with such technology.” They certainly offer a unique skillset.
Project Vigilant
In recent times, super recognisers have played an interesting role in Project Vigilant – a proactive policing tactic developed by Thames Valley Police to tackle VAWG.
Vigilant was first launched in 2019, following a rise in non-domestic sexual offences in Oxford. Its aim is to identify people showing signs of potentially predatory behaviour – such as sexual harassment – and intervene pre-emptively, to prevent sexual offences from taking place.
To achieve this, Thames Valley Police deploys uniformed and plainclothes officers outside of nightclubs, pubs and bars, to identify individuals displaying this kind of behaviour. This is underpinned by the specialist training all Project Vigilant officers receive – so far, 150 Thames Valley staff have been through this process.
Rather than reactive policing, arresting those who have already committed sexual offences, this model is explicitly preventative. Uniformed Project Vigilant officers actively engage with individuals behaving in this way, explaining the behaviour that has raised concerns, and only rarely leading to immediate arrest.
Vigilant also uses a diversion scheme, with officers able to refer those stopped onto a bespoke intervention programme run by charity Circles South East, an organisation that seeks to boost desistance among sexual offenders.
The programme has already borne some success. Thames Valley Police claims that the pilot version of Project Vigilant (back in 2019) led to a 30% cut in sexual offences of this type. Fuller results are expected to emerge from a separate, three-year academic evaluation, and several other forces have already decided to copy this model.
There is, though, another crucial feature to Project Vigilant: all stops are recorded on officers’ body-worn video. And this is where the super recognisers come into play.
Super vigilance
Thames Valley Police has a three-step process to recruiting its super recognisers. First, officers are encouraged to complete the Greenwich test – a tool designed to identify those with exceptional facial recognition ability.
If they score nine or above, they are invited onto the next stage: an extended assessment that delves further into their skills. Pass this too and there’s a week-long training scheme, designed to show how this skill can be applied to the day-to-day work of fighting crime. These certified super recognisers can then be drafted to support Project Vigilant.
With every Vigilant stop recorded on body-worn video, Thames Valley Police are able to collect a database of suspect images. Super recognisers familiarise themselves with these images before going out on patrol with Vigilant; this means that – when deployed – they can identify individuals already known to police as engaging in potentially predatory behaviour.
“
In addition to creating safer spaces, we have sought to increase the knowledge, understanding and operational skillset for officers across the Thames Valley in how to spot predatory behaviour by developing a bespoke training package.”
DI Tina Wallace, Thames Valley Police
The same goes for CCTV. Super recognisers can review CCTV images and spot potential offenders who have had previous contact with Vigilant officers. Pulling together the intelligence picture is the overriding focus of this work – a more holistic approach to taking on the sexual harassment behaviour that defines far too much of the night-time economy.
However, as the force points out, their super recognisers play a role in other contexts too. This includes deployment at large-scale events such as football matches or the annual Reading festival, as well as day-to-day support for policing. In addition the ‘super recogniser mailbox’ also picks up images of unknown people from across the force, enabling these individuals to be identified much more quickly.
Detective Inspector Tina Wallace, part of the Specialist Operations team at Thames Valley Police, emphasises the role of super recognisers: “Project Vigilant is a proactive initiative to prevent sexual offending in the night-time economy… super recognisers form part of our Project Vigilant cadre, this includes both police officers and police staff, whether they are deployed physically within the night-time economy or CCTV based.”
She added: “As well as being adept at recognising faces, our super recognisers have given other officers and staff the confidence to take the Greenwich test. This test has led Thames Valley Police officers and staff to being assessed as super matchers. These are individuals who are super skilled at recognising a face from a previous photograph they have seen.”
There are also broader benefits of Project Vigilant across the force’s work, as DI Wallace explains: “In addition to creating safer spaces, we have sought to increase the knowledge, understanding and operational skillset for officers across the Thames Valley in how to spot predatory behaviour by developing a bespoke training package. The skills taught are transferrable to other areas of policing in a plain clothes or uniform capacity.”
The future of policing
Project Vigilant is an interesting case study of how policing can be done differently. Proactive not reactive, prevention-led rather than arrest-first, it’s a fresh approach to a deep-rooted issue.
To quote evidence submitted to the House of Commons Women and Equalities Committee in 2018: “Harassment isn’t just expected, it’s accepted. It is a cultural norm and women are no longer surprised to be hassled, harassed or assaulted. Many never report it, they just see it as a normal part of a night out.” More innovative approaches like Vigilant might be part of the solution to this problem.
Project Vigilant also raises other questions about the role of technology. In an era where AI is the watchword and ChatGPT is now a household name, the role of super recognisers is revealing.
Perhaps in policing, as in other sectors, the answer is human and machine working together. Can live facial recognition be augmented with a (super)human recogniser element? Or does this create as many ethical problems as it solves? Only time will tell.
It is vital for those in command of law enforcement officers, border guards, security at a venue or a CCTV control room to know which of their team suffers from prosopagnosia (face blindness) or which of them are super recognisers. This is especially the case if your role is to spot criminals or confirm identities. The Association can assist to identify these traits.
From the article: “…sufferers include…actor Brad Pitt, 59, who said prosopagnosia means “so many people hate me because they think I’m disrespecting them”. According to a study published in the journal Cortex in February, millions of us may have prosopagnosia without even knowing the condition exists.
Previously thought to affect between 2 -2.5 per cent of the global population, researchers at Harvard Medical School (HMS) and the VA Boston Healthcare System, who administered an online questionnaire and tests to 3,341 individuals, found prosopagnosia might actually affect as many as one in 33 people – 3.03 per cent.
…At one end are “super recognisers”, who can instantly recall the face of a person they haven’t seen for decades. Then there are those of us who might struggle to put a name to the face of an office colleague outside the context of work. But at the other end, says Dr Shah, are people with extreme prosopagnosia, “who can’t recognise themselves in photos”.”
See the article here:
https://www.telegraph.co.uk/health-fitness/mind/scientific-reason-remember-faces-names-prospangnosia/
Research conducted by the University of New South Wales and other universities has further demonstrated how Super Recognisers can assist law enforcement and border control agencies - especially operationally, where fast decisions are required. The strength and weaknesses of the various methods of identification - SRs, trained forensic examiners and automated systems (referred to a DNNs) and how combining all methods produces excellent results. From the article:
We found that forensic examiners, super-recognizers and DNNs all achieve high accuracy, but each have distinct strengths and weaknesses which make them suited to different real-world face identification roles. Super-recognizers can make decisions quickly and rarely miss targets. Super-recognizers are therefore most suited to time-critical roles where the priority is to avoid false negative errors (misses) in the interests of public safety, such as border control, surveillance, searching for a face in a crowd, and reviewing the output of automated database searches. In these roles, false positive errors—which super-recognizers are prone to making—can often be eliminated quickly by further investigation
NOTE - in the SR system developed by Mike Neville, identifications are researched and peer reviewed. All are also subject to the rigorous identification procedures demanded by the law (in UK, PACE 1984 Code D, Part B).
https://www.nature.com/articles/s41598-023-28632-x
In this fascinating article, it is clearly shown that a computer only sees a face - it does not know if it is real or not. In this study, an image of a face on a T-shirt was used to trick the computer programme. The Chief Executive of the Association has also witnessed this issue when working with Artificial Intelligence. When a test subject wore a T-shirt depicting the Beatles, the system could not function as it could only deal with one face at a time, but it was seemingly confronted with five - the real face, together with images of John, Paul, George and Ringo!
This is another reason to utilise the skills of HUMAN Super Recognisers to enhance and improve any automated facial recognition. The article can be found at this link:
The ASR is the professional body representing those who possess the super recogniser skills around the globe. Via our associate scheme, you will be able to access the latest developments in the science of super recognition, browse job opportunities and contact other practitioners from around the globe who share your unique skills and abilities. After a certain period of years in practice, you may even be invited to become a Fellow of the Association.
Have you got an amazing memory for human faces? Your skill could help law enforcement, security companies, Armed Forces or other organisations. You may just want to use your skill for fun OR are simply interested in Super Recognition.
There is NO need to have a background in security, the military or law enforcement. Super Recognisers come from a wide variety of backgrounds.
Join Contact Privacy Policy Cookie Policy © 2023 Association of Super Recognisers
