How France Is Using AI To Check If People Are Wearing Virus Masks in Public

The concept is to identify potential future hotspots by cross-referencing current behavioral data with historical experience. JL

James Vincent reports in The Verge:

France is integrating new AI tools into security cameras in the Paris metro system to check whether passengers are wearing face masks. The goal is not to identify or punish individuals who don’t wear masks, but to generate anonymous statistical data that will help anticipate future outbreaks of COVID-19. It’s likely it will become a staple of urban life. As countries begin to weigh the economic damage of a lockdown against the loss of life caused by more COVID-19 infections, greater pressure will be put on mitigating measures like mandatory masks.

France is integrating new AI tools into security cameras in the Paris metro system to check whether passengers are wearing face masks.

The software, which has already been deployed elsewhere in the country, began a three-month trial in the central Chatelet-Les Halles station of Paris this week, reports Bloomberg. French startup DatakaLab, which created the program, says the goal is not to identify or punish individuals who don’t wear masks, but to generate anonymous statistical data that will help authorities anticipate future outbreaks of COVID-19.

“We are just measuring this one objective,” DatakaLab CEO Xavier Fischer told The Verge. “The goal is just to publish statistics of how many people are wearing masks every day.”

The pilot is one of a number of measures cities around the world are introducing as they begin to ease lockdown measures and allow people to return to work. Although France, like the US, initially discouraged citizens from wearing masks, the country has now made them mandatory on public transport. It’s even considering introducing fines of €135 ($145) for anyone found not wearing a mask on the subway, trains, buses, or taxis.

Example statistical data generated by DatakaLab’s software.

 Screenshot: DatakaLab

The introduction of AI software to monitor and possibly enforce these measures will be closely watched. The spread of AI-powered surveillance and facial recognition software in China has worried many privacy advocates in the West, but the pandemic is an immediate threat that governments may feel takes priority over dangers to individual privacy.

DatakaLab, though, insists its software is privacy-conscious and compliant with the EU’s General Data Protection Regulation (GDPR). The company has sold AI-powered video analytics for several years, using the technology to generate data for shops and malls about the demographics of their customers. “We never sell for security purposes,” says Fischer. “And that is a condition in all our sales contracts: you can’t use this data for surveillance.”

The software is lightweight enough to work on location wherever installed, meaning no data is ever sent to the cloud or to DatakaLab’s offices. Instead, the software generates statistics about how many individuals are seen wearing masks in 15-minute intervals.

The company has already integrated the software into buses in the French city of Cannes in the south of the country. It added small CPUs to existing CCTV cameras installed in buses, which process the video in real time. When the bus returns to the depot at night, it connects to Wi-Fi and sends the data on to the local transport authorities. “Then if we say, for example, that 74 percent of people were wearing a mask in this location, then the mayor will understand where they need to deliver more resources,” says Fischer.

TECHNOLOGY LIKE DATAKALAB’S WILL LIKELY BE COMMON IN FUTURE

Although technology like DatakaLab’s is only being tested right now, it’s likely it will become a staple of urban life in the near future. As countries begin to weigh the economic damage of a lockdown against the loss of life caused by more COVID-19 infections, greater pressure will be put on mitigating measures like mandatory masks. In countries in the West where mask-wearing is more unfamiliar, software like DatakaLab’s can help authorities understand whether their messaging is convincing the public.

Fischer says that although the pandemic has certainly created new uses cases for AI, it doesn’t mean that countries like France need to abandon their values of privacy and embrace invasive surveillance software. “We respect the rules of Europe,” says Fischer. “This technology is very useful but can be very dangerous ... [But] we have our values and they are part of our company.”