- The Chinese city of Guiyang said its facial-recognition software allows police to identify and arrest suspects in as little as two minutes.
- Cameras are placed in more than 10,000 public areas and send surveillance feeds to police in real-time, identifying individuals, their family members and where they have been in the last week.
- A BBC reporter tested the system in December and was identified and captured within seven minutes.
- China has hit back at concerns over privacy, saying that there are no problems because the data is collected in “public places.”
An expansive facial recognition network in the southwestern Chinese city of Guiyang have reportedly enabled police to detect and apprehend criminals in as little as two minutes.
The surveillance cameras are set up in more than 10,000 public places across the city, which is larger than the city of Perth, reported the state-run newspaper Global Times this week. The cameras stream real-time footage back to a huge LED screen – 22 metres (24 yards) long and 5 metres (5 years) tall – which is monitored by police.
The system, which has a 90% accuracy rate, simultaneously checks faces against a nationwide database and can almost immediately provide a person’s name, age, gender, ethnicity, as well as information including family members, people they regularly meet, and places they have recently been.
“The screen captures images of any suspects once they arrive in Guiyang, and then automatically reports them to the police command center which immediately allocates nearby police to make the arrest. The whole process from detecting to apprehending suspects usually takes less than two minutes,” Global Times reported, citing a press release from local police.
The development is part of Skynet, a nationwide monitoring programme launched in 2005 to increase the use and capabilities of surveillance cameras.
In Guiyan last year, facial-recognition technology led to the apprehension of 375 suspects, including 39 fugitives, according to police.
“Guiyang has ‘Skynet’ everywhere. No matter where you go, there are eyes on you,” Li Bin, an official at the Guiyang Public Security Bureau, said in a press release.
A BBC reporter put the Guiyang system to the test in December last year. In a city of 4 million people, it took only seven minutes for the reporter to be identified and arrested.
But as quick as Guiyang’s police response times are, the number of cameras is lagging behind Beijing, which achieved 100% coverage of the city in 2015.
Currently, there are 170 million surveillance cameras in China and by 2020, the country hopes to have 570 million – that’s nearly one camera for every two citizens.
Police are also developing AI-powered systems that, aside from recognising faces, can identify people from their repeated behaviours or even gait. Officials want to use this information to predict crime before it happens.
China hits back at claims its surveillance infringes citizen’s rights
China has begun hitting back at human rights concerns over its expanding surveillance and facial recognition technologies.
“Some mainstream Western media outlets define the system as a tool for massive surveillance and suppressing human rights,” an editorial in Global Times said last year. “This is the kind of fake news that US President Donald Trump often slams.”
Part of the issue is that the government in China defines a person’s right to privacy differently than Western countries.
The Global Times reported the system in Guiyang, like those across China, “does not infringe on people’s privacy and human rights, as facial information is only collected in public places.”
Its this approach to privacy that, until recently, has been a huge driver in China’s soaring AI sector, according to a report out this month from Oxford University.
Lax privacy protections have allowed Chinese companies to access masses of data to develop AI and technologies like facial recognition technologies because “sharing among government agencies and companies is common,” the report said.
Watch the BBC test here:
— BBC News (World) (@BBCWorld) December 10, 2017