Belgrade 'Safe City' video surveillance 

Released: 2019

Can you improve this page?
Share your insights with us

Serbia's Ministry of Interior has unlawfully installed over 1,200 Huawei real-time facial recognition CCTV cameras across Belgrade, according to reports in WIRED and other media outlets.

Piloted in 2018 and launched in 2019 with the stated aim of installing 800 cameras with facial and vehicle licence plate recognition, the expansion marks a major expansion of the Chinese company's 'Safe City' security programme with the Serbian city. 


The expansion has led to accusations of covert surveillance and opacity by human rights and privacy activists, EU members of parliament, and members of the general public. It also resulted in #hiljadekamera, a grassroots campaign led by Serbian digital rights organisation SHARE Foundation.

The system was the subject of a review by Serbia's data privacy commissioner, who in 2019 found that the Ministry of Interior's Data Protection Impact Assessment had not been conducted in line with Serbian national law, arguing 'there is no legal basis for [its]  implementation'.

At the time, Serbia had no data privacy law. The country's Personal Data Protection Law went into effect in August 2019.


According to Privacy International, Serbia's Ministry of Interior reputedly insists all information about the system is 'confidential' and that it holds no information on locations and crime rate analysis for the city.

The Ministry is also said (pdf) to have been routinely blocking freedom of information (FOIA) requests for information on the location and workings of the cameras and system.

Meantime, a Huawei case study about a secret 2017 test of the system in Belgrade was removed from its website.

Operator: Ministry of Interior
Developer: Huawei
Country: Serbia
Sector: Govt - interior; Govt - police
Purpose: Strengthen law enforcement
Technology: Facial recognition
Issue: Privacy; Surveillance
Transparency: Governance; Black box; Privacy; Legal


Research, advocacy

News, commentary, analysis