Facial recognition cameras installed by a supermarket chain to tackle shoplifting disproportionately target people in poorer areas, according to a privacy rights group.
Southern Co-op, which uses Facewatch live recognition cameras in 34 branches, typically has shops in richer-than-average neighbourhoods. But just five of the stores in which it uses Facewatch are in the richest third of neighbourhoods in England, while 14 are in the poorest.
Professor Pete Fussey, director of the Centre for Research into Information, Surveillance and Privacy at University of Essex, said the findings raised concerns about the targeting of people on the “margins of society”. “This is another example of the many ways in which surveillance is more intensely focused on minorities and those who are disadvantaged socioeconomically,” he said.
Southern Co-op, which is separate from the national Co-op chain, said it did not consider how deprived an area was, or other demographic information, when deploying the technology. The cameras were in locations considered high-risk based on “our crime data and stock loss reports”.
Fussey said police and businesses often claimed decisions on where to use facial recognition were an “objective judgment” based on crime levels. “But the issue is that it’s not clear at all,” he said. “We’re going to find crime in the areas we look at most heavily. And we look most heavily at suspect populations, those who are seen as threatening or outsiders, and the urban poor.”
Analysis by the privacy group Big Brother Watch used data supplied by Southern Co-op cross-referenced with the England-wide 2019 Indices of Multiple Deprivation. The index ranks more than 30,000 neighbourhoods across seven areas, including income and employment, to calculate their relative deprivation – with the number one ranked neighbourhood being the most deprived.
An average Southern Co-op store is in an area ranking at 19,835 out of 32,844, putting it in the best-off third of neighbourhoods. But supermarkets where facial recognition is deployed are in neighbourhoods ranked at 14,453 on average, placing them in the most deprived half .
Jake Hurfurt, the head of research at Big Brother Watch, said: “This data shows that AI supermarket surveillance is being directed at poorer communities, who are more likely to suffer excessive invasions of their privacy and unfair treatment as a result.”
Increasingly used by police and private firms, live facial recognition operates in real-time to compare camera feeds with faces on a predetermined watchlist, to identify people of interest. Each time a match is found, the system generates an alert.
Advocates including police chiefs and the government say the technology deters crime and helps identify offenders. But concerns have been raised over a lack of oversight and transparency as its use becomes more mainstream.
On Friday, the House of Lords Justice and Home Affairs committee wrote to the home secretary, James Cleverly, calling on him to urgently address concerns about live facial recognition use by police, which it said lacked “clear legal foundation”.
The committee said there were “no rigorous standards or systems of regulation” in place for monitoring the technology’s use and “no consistency in approaches to training” among police forces. “The committee accepts that live facial recognition may be a valuable tool in apprehending criminals, but it is deeply concerned that its use is being expanded without proper scrutiny and accountability,” it said.
Fussey described use of the technology in Britain as a “wild west”. “There is no proper, robust, meaningful regulation of it. There’s even less in the private sector than there is in the public sector,” he said.
Nick Fisher, the chief executive of Facewatch, said the technology was used to protect stock and staff and to deter crime. “Retailers across the UK use Facewatch in those stores where it is necessary to do so because other crime prevention methods have been tried and failed,” he said. “Their lawful use of Facewatch reduces crime between 30% and 70% in every location it is deployed in.”
A Southern Co-op spokesperson said its use of the technology was “limited and targeted”, and had helped cut crime. It recorded 38,262 incidents in 2022, including “shoplifting, violence and physical and verbal abuse”. Internal figures showed a 34% reduction in violence against staff in 2021 compared with 2020 at shops using the technology.It acknowledged that this was due to “a range of measures we are using to tackle crime, as well as a number of external factors”.
The spokesperson added that Southern Co-op did not maintain a “watchlist” but its facial recognition provider had “lists of individuals of interest who are known and evidenced as having offended, including those who have been banned/excluded”.
In October, human rights groups including Liberty and Amnesty International wrote to leading UK retailers urging them to quit a police-led scheme that aims to use facial recognition to tackle shoplifting. In a letter to companies including John Lewis, Tesco, Boots and M&S about the “Pegasus” initiative, the charities said: “Facial recognition technology notoriously misidentifies people of colour, women and LGBTQ+ people, meaning that already marginalised groups are more likely to be subject to an invasive stop by police, or at increased risk of physical surveillance, monitoring and harassment by workers in your stores.”
A government spokesperson said facial recognition a “sound legal basis” and had helped police catch “a large number of serious criminals” including people wanted for murder and sexual offences. “The police can only use facial recognition for a policing purpose, where necessary, proportionate and fair, in line with data protection and human rights laws,” they said.
ONS figures last week put shoplifting at record levels, with more than 402,000 offences in the year to September 2023, up from 304,459 the year before.
