Tag: facialrecognition

CCTV facial recognition and the Hong Kong protests

Artificial Intelligence, Big Data, Digital Systems Technology, Internet of Things (IoT)

August 24 marked another long day of clashes in Hong Kong between police and protestors.

The nearly three-month long protests generally choose their location based on certain problematic aspects of Hong Kong’s relationship with China. 

That day, the focus opposed police digital surveillance tactics, specifically the city’s installation of ‘smart lampposts’; street lights that come equipped with cameras and sensors. Hong Kong authorities insist the lampposts are only used to monitor weather, traffic, and air quality. 

Police and protestors on Aug 24

After riot police forcibly dispersed an initial gathering of tens of thousands of Hong Kong citizens, groups of protestors broke off and regrouped in different neighborhoods. Several smart lamps were pulled down, dismantled, and occasionally kicked. Police fired projectiles, released tear gas, and wielded batons and riot gear while charging lines of protestors.

Facial recognition already common

Various facial recognition technologies and devices are already common across Hong Kong. They unlock mobile devices, confirm identities at ATMS, and identify individuals in photos based on their social media profile.

Suspicions of increased mainland monitoring

The renewed concern of protestors directly relates to the catalyst for the entire protest movement: the extradition bill that sought to allow Hong Kong citizens to be tried by the mainland Chinese judiciary system.

The general public simply does not trust the new cameras, many of them developed and sold by Chinese companies, to innocuously report weather and traffic data. 

China obviously wants the ability to extradite Hong Kong citizens for crimes committed outside of the mainland. A video surveillance network in which Chinese authorities can monitor individual Hong Kong citizens is a clear path towards building cases against political targets that could facilitate a case for extradition. 

Darwin council claims new CCTV cameras “will not use facial recognition”

Internet of Things (IoT)

Darwin council insists that the new set of cameras in the Central Business District will not employ facial recognition until the technology is more carefully considered.

The 138 cameras were installed as part of the Switching On Darwin project, which also provides public wifi, new lighting, and traffic sensors. Although the cameras do have facial recognition hardware in them, the city does not plan to utilize it at this time.

However, council general manager Josh Sattler stated that although the city will not employ the technology, “requests coming from a state or federal agency” will “need to be complied with.” 

Install first, question later

Of the 81 grants offered by Australia’s Smart Cities and Suburbs program, only the Darwin proposal included funding for facial recognition technology. The Darwin council’s addition of the technology was done without public consultation. An estimated six month study by an outside consultant agency began in August.

Concerns with facial recognition

Many question the council’s decision to include a technology that will supposedly not be employed. Julia Powles, researcher and associate professor at the University of Western Australia, questions “the real reason why [the technologies] were sought in the first place”, adding “it’s like buying a supercomputer and then saying you’re just going to surf the web and check emails.”

Security of camera feeds also stands out as a major concern. Elise Thomas from the International Cyber Policy Centre, offered a warning along these lines:

“The ways in which a lot of these projects are being implemented at the level of local councils, particularly from a cyber-security perspective, those councils may not necessarily have the resources to ensure that their systems are secure”

Internet connected devices are notoriously vulnerable to security breaches. Any gateway into a network-connected device can open doors to the entire network. Depending on the type of access provided, compromised devices can be mined for data, or even remotely controlled by outside parties.