Intention: Open doors in the nursing home
Since 2019 uses Sevagram, a care institution with various nursing homes in Limburg, the 'freedom', unless' principle. Residents are allowed to move freely within and around the location, provided their own safety and that of others can be ensured. Residents wear a wrist transmitter that sends a notification to the healthcare professional when they walk through the wrong door.
‘How nice would it be if residents were no longer stigmatized by a wrist transmitter?Spy, Sevagram wondered. The healthcare institution contacted software developer Theo Breuers, who has experience with facial recognition systems in, among other things, football stadiums. He sees opportunities for nursing homes.
Breuers' system issues a warning to nursing staff when a resident enters or leaves a certain room. The system stores from registered persons (with permission) some facial features ("vectors") on. When the camera detects someone, the system compares the characteristics of that person with the registered characteristics. It is then determined what rights the person with those characteristics has, without the system knowing who the person is. No face can be reconstructed from the attributes and no photos will be saved. Privacy by design, would you say.
“How nice would it be if residents were no longer stigmatized by a wristband?”
Approach: A system based on facial recognition, tested against the GDPR
In collaboration with Sevagram, Theo Breuers immediately started developing the system. He sought contact with partners who, among other things, open- and develop locking techniques because his company Compo does not own them. Since many care home systems are not GDPR-proof, Breuers made sure that his system could function on its own.
Breuers knew from the start that privacy legislation could create barriers. After all, you can't just store data from casual passers-by. People must be able to actively give their consent before they are portrayed, and agree to the use of their data. That is why Breuers had the system checked against the GDPR. "That seemed okay. A nursing home is not a public building and everyone who enters the building, could theoretically sign something to give permission. We also didn't hang cameras on the outside of the building, so you don't have any 'by-catch' from chance passers-by.'
“The law is misinterpreted when it comes to facial recognition.”
Result: Blocked by a strict interpretation of privacy law
Even before a pilot could be started, a second check with external parties revealed that the system did not comply with privacy legislation. It was not allowed to be used. "The law is misinterpreted on facial recognition",’ thinks Theo Breuers. According to him, it is important to distinguish between security systems and facial recognition. Security cameras record everything and everyone that passes by, while Breuers' system has no so-called by-catch.
To get around this problem, Breurers should build a second system for everyone who visits the nursing home, with the exception of the clients, based on, for example, barcodes or living circles. ‘We have decided not to do this because we still have a stigmatizing effect. That would bring us back to square one.”
"This problem will soon also occur in other places where facial recognition is used.",’ he suspects. It seems that the translation of the GDPR into law- and regulations have not taken into account future developments. In this case, new developments make it possible to observe people, without violating their privacy.