Last month, the Mayor of London’s office quietly signed off on the purchase of ‘Retrospective Facial Recognition’ (RFR) systems that processes existing images accessed by the Metropolitan Police from a variety of sources and matches them against the force’s internal database.

Under the four-year agreement, Japanese tech firm NEC Corporation will supply the RFR systems – which will afford the Met Police “opportunities... not previously available to support the detection and matching of faces” and enable the force to “effectively exploit” investigative opportunities from the rapid growth of “image data sources”.

Although the system is due to go live by year’s end, the country’s data watchdog – the Information Commissioner’s Office (ICO) – told Wired UK that it had not published any official guidance on the use of RFR systems and warned that “public trust” would be lost if police forces using the tech did not complete “crucial steps” to comply with data protection laws “before, during and after its use”.

Among these compliance requirements is the completion of a Data Protection Impact Assessment (DPIA) prior to processing personal data. According to the tech news outlet, the Met had not yet submitted a DPIA when the Deputy Mayor for Policing and Crime approved the RFR purchase – despite the agreement proposal noting that the system will “ensure a privacy by design approach”.

But an unnamed Met spokesperson said it needed to identify a vendor before publishing the DPIA. They added that the use of images would be “subject [to] a carefully implemented framework” regarding “expectations of privacy” but did not share details of what this framework entails.

In addition, the London Policing Ethics Panel – an independent oversight body created by the mayor’s office – had reportedly not reviewed or advised the Met on the use of RFR prior to the signing of the agreement. The panel has not commented on the purchase as yet.

An unidentified mayor’s office spokesperson claimed that the tech would shorten the time needed to identify suspects and help tackle crime rates in the capital, but admitted that the Met needed to be “proportionate and transparent” with its use to “retain the trust of all Londoners”.

Last week, London Mayor Sadiq Khan released the ‘Emerging Technology Charter’, which outlines a number of guidelines on the use of new data-enabled technology in the city. Although exempting law enforcement bodies, the document noted that tech like facial recognition systems “should not be deployed” if they did not meet the “very high bar” set by the ICO for using biometric data.

According to a report by police watchdog HM Inspectorate of Constabulary and Fire and Rescue Services, RFR technology is currently used by six police forces in England and Wales. However, Wired UK noted that there was “sparse” publicly available information and scrutiny on how these police forces use it – as opposed to the spotlight around ‘Live Facial Recognition’ (LFR) technology, which is used by the Met and other forces to scan faces in real-time and match them to a “watchlist”.

The Met has continued to deploy LFR tech despite a House of Commons committee recommending against its use in July 2019. In its purchase proposal, the force repeatedly states that RFR use is “very different” to LFR. The Met reportedly sources both products from the NEC Corporation, which declined to comment.

However, Daragh Murray, a University of Essex lecturer who has reviewed the Met’s use of facial recognition tech, told the outlet that the two systems had the potential to be “massively invasive” and could be “strikingly similar” depending on how they are deployed.