Auror Subject
Recognition

The FRT providers we integrate with are responsible for developing, maintaining and improving their technology. Retailer data is not used to train any algorithm and any information stored in Auror is never used to train FRT models.

Auror provides retailers the tools to determine how they manage information relating to minors. Many jurisdictions and retailers’ own operating policies require information about minors to be treated sensitively. The unfortunate reality for frontline retail workers dealing with risk every day is that a minor with a weapon can be just as threatening as an adult with a weapon. Retailers can choose to be alerted for specific levels of risk related to minors or exclude minors from biometric processing altogether. Retailers can control their own response based on what is proportionate for their unique environment and risk setting.

Leading FRT is incredibly accurate and we only work with providers that are independently benchmarked and deliver world-class results validated by accredited bodies, such as the US NIST.

Accuracy of FRT is only part of the equation when it comes to bias and misidentification. Unlike other systems, Auror prohibits the specific collection of sensitive information such as ethnicity, race, political or religious affiliation or sexual orientation. Therefore, Subject Recognition cannot be used to categorize people into groups based on these demographics, and prevents profiling and targeted advertising based on these groups. It cannot be used for loyalty programs or other tracking purposes.

The FRT relies on a number of factors including the reference image quality and the setting of confidence thresholds. While retailers choose these thresholds and are responsible for the reference images they upload, Auror sets minimum thresholds to guide effective and responsible use. Retailers can configure these levels to balance the rate of false positives and false negatives, to both protect their store teams and reduce misidentification.

Ultimately, FRT is a tool that supports human decision-making, and should not be used to make automated decisions for the retailer. Subject Recognition is designed to have trained individuals involved at each key decision-making stage and provide crucial real-time context to support their response when a potential threat enters the store.

No record of any individuals who are not on the POI list is ever retained. For good, honest shoppers, their temporary biometric template is instantly discarded and never stored. Where there is a match, the detection image is sent to Auror to generate an alert, alongside the reference image (a still photo), for a trained retail employee to review. No biometric data (match or non-match) is stored in the cloud. 

In all cases, retailers determine the length of data retention with system-defined limits to ensure data is never kept for longer than necessary.

Retailers authorize a limited number of specific roles in their organization to have access to the Subject Recognition. Typically retailers restrict access to a smaller, trained group within their organization to review the POI list, and accept or decline suggested enrollments.

No. Subject Recognition has been configured in such a way to allow a retailer to keep their own information private to themselves. A retailer can configure their POI list by store, region, or organization-wide rather than a ‘master’ list controlled by a third party. Through an end-to-end workflow, the system allows retailers to benefit from information about past offending across their store networks, while enhancing data minimization, reducing data duplication by individual stores.

No. Subject Recognition can only be used for crime prevention and safety purposes. It has no capability to build general shopper profiles, analyze purchasing behavior, track, categorize, or target for marketing purposes.

Retailers own and manage their own information. Subject Recognition has been configured in such a way to allow a retailer to keep their own information private. Auror does not on-sell data to third parties or use data for its own purposes.

‍

Retailers use Auror’s core software platform to document crimes and in-store events after they happen at their sites. Profiles linked to events that meet the retailer’s criteria for serious and high harm offending are suggested by the Subject Recognition module for enrollment to that retailer’s own POI list. 

An authorized, trained retailer employee can review these suggestions based on their own internal policies. Retailers cannot arbitrarily enroll anyone to the POI list (i.e. taking images from online) and the post-enrollment auditing functionality allows retailers to ensure quality control of their list.

If a retailer changes any information that means a profile no longer meets the criteria for enrollment, then that profile automatically falls off the list.

Subject Recognition is deliberately built as a retailer tool. It exists to help retailers identify high risk, serious offenders that have previously offended in their stores, and support the decision-making of their frontline teams according to risk. It doesn't automatically alert police, and police don't have access to it.

There are a few reasons for that. Retail crime is extremely high volume and automatically alerting police to matches, in addition to the crime reports they already receive, would overwhelm agencies with notifications they might not be able to act on, and could risk giving retailers a false sense of security that a report to police has been made. We believe the right response is for the retailer to assess the situation and decide, based on risk, whether police involvement is needed and proportionate to the risk.

We work with law enforcement agencies around the world, and we know any FRT used in a policing context needs to be owned, governed, and independently verified by police themselves - not via a retailer or third-party’s system.

Depending on the jurisdiction, retailers are required to and are responsible for clearly communicating their use of FRT through appropriate signage. Auror encourages retailers to use signage that links to privacy policies for both the retailer and Auror. This provides customers with the opportunity to choose whether they want to enter the store, go to a different store, or use an alternative method such as online shopping. The purpose of Subject Recognition is to give frontline retail workers early notification of individuals previously involved in serious or high harm offending in their stores.

Concerns around FRT are typically based on its potential use for profiling, tracking, behavioral monitoring, marketing purposes, loyalty programs, or scraping images from open sources. Retailers cannot in any way use Subject Recognition for those purposes. 

Unlike many other solutions, retailers own and manage their information and this is not shared with other retailers or law enforcement. Auror prohibits the specific collection of sensitive information such as ethnicity, race, political or religious affiliation or sexual orientation. To be enrolled to the POI list, an individual must already have previously offended at a retailer’s store and meet the retailer’s criteria for serious, high harm or high loss offending.

Built-in integrity tools provide retailers with full auditability. A human is always in the loop on key decisions and no action is automated. Biometric processing takes place in-store and not in the cloud. If there is no match, that biometric data is discarded immediately.

Retailers using this technology will have signage in place upon entry as notice of their use of FRT, generally linking to their own privacy policy.

On entry the FRT detects a face and creates a temporary biometric template which is cross-checked instantly against a POI list of individuals that have been involved in serious offending.

This happens instantly on the camera itself or an in-store server. If there is no match, that biometric data is discarded immediately and never stored.

If the system suggests a match, an alert is sent via Auror’s cloud to an authorized, trained retail employee with the ability to view context about the POI’s previously recorded serious offending. If they reject the suggested match, the biometric data and detection image is discarded.

If the match is confirmed, the employee can decide what the best response is, according to that retailer’s training and internal policies. They are supported by information from previous offending within Auror (i.e. the individual is known to carry a weapon). The response can be as simple as engaging politely at the store entrance, not approaching or calling police. Any outcome and follow up is also captured, so a retailer can see the effectiveness of their system.

Retail crime is becoming more brazen, organized and violent, and the obligation for retailers to keep their frontline workers safe has become clear. FRT capability and accuracy has significantly improved in recent years, and will be a key tool in reducing violence in stores alongside other recognition technology like license plate recognition (LPR).

While Auror does not build or own FRT algorithms, it applies the secure data governance layer, end-to-end workflows, and safeguards to ensure it is used effectively and responsibly for specific purposes. This keeps retailers safe from non-compliance, protects honest shoppers from privacy breaches, while holding high harm offenders accountable.

FRT is already used in everyday environments such as airports, hotels and casinos, stadiums and in some retail settings. It offers retailers the fastest and most accurate way to identify high-harm POIs and respond according to risk.