Watching Eyes: Face Tech's Double Edge
Facial recognition offers enhanced security but poses serious privacy threats and bias risks, seen in global applications.
Topics: technology, politics, individual and values
Example
China has built one of the world's largest camera networks, including the 'Skynet' project. In December 2017, the BBC tested it in the city of Guiyang: police registered reporter John Sudworth's photo and found him in seven minutes. In April 2019, the New York Times reported that police were using facial recognition to identify Uyghurs by their appearance and track their movements, with one city's system checking whether residents were Uyghurs 500,000 times in a month.
Key facts
- In a December 2017 BBC test in Guiyang, police using China's camera network found a registered reporter in seven minutes.
- The New York Times reported in April 2019 that Chinese police used facial recognition to profile Uyghurs, with one city screening residents 500,000 times in a month.
Limitations
- The BBC test was staged with police cooperation and a pre-registered photo, so it shows capability under ideal conditions; Chinese authorities deny targeting Uyghurs.
Evaluations
technology evaluation
Support
Finding a person among a city's millions in seven minutes shows how cameras and AI together make tracking cheap and fast.
Counterargument
Facial recognition makes errors, especially with poor lighting and for some groups, so systems are less precise than demonstrations suggest.
Rebuttal
Errors do not make the system harmless: a wrong match can bring police to an innocent person's door.
Additional support
Chinese firms export similar systems to other countries, spreading the technology beyond China.
politics evaluation
Support
The reported profiling of Uyghurs shows how a state can use the same tools for crime control and for monitoring an ethnic group.
Counterargument
China says its surveillance cuts crime and terrorism, and many residents report feeling safer.
Rebuttal
Public safety is a real aim, but without independent courts or a free press there is no check on misuse against minorities.
Additional support
The EU's AI Act bans most real-time facial recognition in public by police, showing democracies drawing different lines.
individual and values evaluation
Support
Being identifiable anywhere chills freedom: people may avoid protests, mosques or meetings if they know they are watched.
Counterargument
Some argue that people with nothing to hide have nothing to fear from cameras in public places.
Rebuttal
The profiling of a whole ethnic group shows that harm can fall on people who have done nothing, simply because of who they are.
Additional support
Singapore uses cameras widely but with laws on data use, showing societies balance safety and privacy differently.
Sources
- TechCrunch (2017-12-13): China's CCTV surveillance network took just 7 minutes to capture BBC reporter
- Hong Kong Free Press (reporting The New York Times) (2019-04-16): Authorities using facial recognition tech to track Uighur Muslims across China - report
More GP guides for this example
Topic guides
- Society and Culture GP Examples
- Politics and Governance GP Examples
- Science and Technology GP Examples