Auror launches facial recognition tool in UK stores

Auror has launched its Subject Recognition facial recognition system in UK stores. The tool is designed to alert retail staff to known high-harm offenders.

The rollout brings live facial recognition into a retail crime platform already used by stores to record and track incidents after they occur. The new system uses a retailer's existing crime data within Auror's platform and connects with external facial recognition technology providers, while leaving control of watchlists with each retailer.

The move comes as retailers and police face sustained concern over violence, theft and organised offending in shops. Auror says its UK retail data shows the top 10 per cent of offenders are responsible for almost 70 per cent of retail crime, rising to more than 80 per cent for some retailers.

According to Auror, violence in UK retail stores has risen by 13 per cent this year, while one in 10 crime events involves violence or weapons. The system is intended to help frontline staff identify repeat high-harm offenders before an incident escalates.

How it works

The system does not create shared lists across retailers, Auror says. Each retailer sets its own thresholds for serious offending and builds watchlists from its own records, while police cannot access those retailer person-of-interest lists.

Non-match data is discarded instantly, and biometric template data from matches is never stored in the cloud, according to Auror. The product also does not allow users to tag people with sensitive characteristics such as race, ethnicity, religion, political affiliation or sexual orientation.

Auror says several points of human review are built into the process and that no key decision is made automatically. Staff who receive an alert are given context about previous offending and decide how to respond.

Nick McDonnell, Global Vice President of Trust and Safety at Auror, set out the company's position on the use of the technology in shops.

"We strongly believe live FRT will have a meaningful impact on store safety in the face of high levels of violent and threatening retail crime, but we know it's critical to get this right for both retailers and the privacy of good, honest shoppers.

"It's great to partner with leading UK retailers to help keep their stores safe using this technology. Retailers know how important it is to retain responsibility and control of their own information, and to know where the data they collect about crime ends up," McDonnell said.

He drew a distinction between Auror's approach and broader concerns about centralised databases or uncontrolled data-sharing between organisations.

"With Subject Recognition, we're not talking about some dystopian future of 'national master watchlists' or sharing retailer information with third parties to unknown ends. It gives retailers control and comfort in their compliance position by not allowing their watchlists to be merged with other organisations or shared far and wide.

"This is a human-centred tool that simply helps frontline colleagues make better decisions based on a retailer's information about past offending. It gives them context on whether an individual has been violent or used a weapon before, and they can decide how they want to respond. It's a very human approach and isn't making any key decisions for them, or automatically flooding police with alerts.

"The right safeguards and data governance are critical because if you don't respect data accountability, it can ruin the promise of this important safety tech for the whole sector for years to come. As a sector, it's so important to get the tech right, because it will be game-changing for public safety in retail settings," McDonnell said.

Pilot findings

Auror cited pilot results from other markets where retailers already use Subject Recognition. The trials showed a marked fall in high-harm incidents, with almost all verified alerts followed by no further safety incident.

The company linked those results to the accuracy of alerts, the judgement of frontline teams and the deterrent effect of visible use of facial recognition in stores.

"It's the effectiveness of the alerting, the flexibility of the frontline teams to respond appropriately, and the deterrence of offenders because FRT is used in stores that combine to drive these outcomes," McDonnell said.

The launch adds to a wider debate in the UK over how live facial recognition should be used in public-facing settings, particularly where businesses are balancing worker safety against privacy, proportionality and data governance requirements. Auror's model relies on retailers using their own historic records of serious offending rather than drawing from pooled industry lists.

McDonnell said the retail sector and police were already making progress against violent and organised shop crime.

"We all know retail crime is brazen, violent and organised, but the dedicated work by retailers and police across the country has been fantastic and genuinely world-leading," McDonnell said.

"The cross-sector collaboration and intense focus on tackling the problem is making a difference not only to the frontline retail colleagues who serve their communities every day, but also to everyday shoppers and families who enjoy the vibrancy of the UK's high streets," he said.