Use facial recognition cameras to identify people refusing to wear masks, says former commissioner

Facial recognition cameras should be used to spot people who refuse to wear face masks, the Government's former surveillance camera commissioner has said.

Tony Porter said the cameras had a high success rate for identifying people whose faces were 50 per cent covered as they would be with a Covid mask or covering.

With the appropriate safeguards, they could be used either to identify an individual without a mask in locations where Covid regulations required their use or on an anonymised basis to check compliance and pinpoint a "blanked out" face without a covering.

Mr Porter, whose post as commissioner has been merged by the Government with the biometric commissioner, said: "Say you have 100 people in a location moving around and there is an absolute requirement for people to wear masks, then this technology can identify a subject without a mask.

"There is no need for a watchlist. If it is just a case of someone not wearing a mask, it is capable of homing in because it can identify who is or is not."

Locations could range from transport hubs and inside trains or buses to shops and hospitals or crowded outdoor places such as stadia and carriageways.

Mr Porter, now the chief privacy adviser to Corsight AI, a technology firm, said there were significant public interest gains from facial recognition technology in a health pandemic provided there were transparent safeguards over privacy and use of data.

He said any operator should commit to five principles including erasing images immediately their use had passed, requiring consent to the use of the technology in "certain circumstances" and the capability to anonymise data such as with the "blank" faces to check on mask use.

There also needed to be a clear public interest for deployment and rigorous standards over the security of any images collected, said Mr Porter, who as commissioner produced a best practice guide for the technology before stepping down in December.

Facial recognition technology has become so sophisticated that it can identify a face from oblique angles, even from 90 degrees above or below. But there are still continued inaccuracies in recognising black and ethnic minority faces.

Videcon, a Yorkshire company, is already promoting AI face recognition technology to stores so that when a customer enters a shop it can recognise if they do not have a mask and alert a member of staff or notify the customer on a screen or via audio of the regulations.

The deployment of the cameras has, however, been controversial, with the court of appeal last year ruling that the way South Wales Police was using them was unlawful. It has been trialled by other forces including Scotland Yard.

The Home Office has said it wants police to use new crime-reducing technology while "maintaining public trust" but is under pressure from watchdogs, including Mr Porter, to develop clearer guidelines.