A new facial recognition system in Co-Op food stores has started an uproar among privacy advocates.
Indeed, the new system is able to detect if someone who enters the store had a past record of theft or anti-social behavior. This was set up in order to protect workers from assaults and attacks by shoplifters. It was first started by the Southern Co-operative, which possesses more than 200 stores in the South of England.
However, privacy groups are concerned about this system. Hence, Privacy International wrote an open letter to the retailer asking about the legality of the technology in stores and if the information was shared with the police.
Yet, the Southern Co-operative assured that the system was GDPR-compliant and that no data was shared with the police. It also stated that the system was set up due to the increase in assaults and violence against the stores’ employees. According to them, it was only set up to protect the staff from past offenders.
The Co-op is, however, not the only store to use facial recognition to recognize thieves. Sainsbury tried something similar with an AI-enable concealment detector in several of its stores. The system was supposed to catch thieves and reported the criminal activity to security staff. The supermarket partnered with ThirdEye, assuring that this system stopped over 5000 attempted thefts.
Having facial recognition systems used on customers in the UK is quite frightening, especially as it was also supposedly reported that the technology suffers from inaccuracy and biases, leading to innocent people being flagged and put on criminal databases.
This kind of tech has yet to prove its accuracy and show that it is not an invasion of privacy. This summer, the use of the technology by British police forces tried to use new technology which was however ruled unlawful shortly after. There are also similar cases all over the US, including in big firms such as Amazon and IBM.
For now, the technology is still being used in Co-op stores but there are no plans to deploy it more widely.