evidoria

← Back to browse

Good practice Imported

Belarus's Kipod AI Video-Analytics Platform — Facial Recognition Used to Track and Prosecute 2020 Protesters

Belarus · Minsk · See the Belarus profile · See the Minsk profile

Evidence: Descriptive / self-reported Top 99% 13/100 · Ask Evidence Copilot about this practice

Belarus's Interior Ministry used the Kipod AI video-analytics platform, built by Synesis, for facial recognition on public/metro cameras; after the 2020 election it reportedly helped identify protesters, including activist Nikolai Dedok. EU sanctioned Synesis in Dec 2020.

2,400+ cases
Protest-linked criminal cases opened in the year after the 2020 election (2020-2021)

Details

Promoter
Ministry of Internal Affairs of the Republic of Belarus (system built by LLC Synesis)
Period
2020–2021
Keywords
public security, surveillance, law enforcement, facial recognition

Context

Belarus's Ministry of Internal Affairs has used the Kipod video-analytics platform, developed by the Minsk-based company Synesis, to apply machine-learning facial recognition to footage from public cameras, including cameras in the Minsk metro system.

Activities

After the disputed August 2020 presidential election triggered mass protests, Belarusian authorities reportedly used the system to identify and track individuals. Privacy International documents the case of activist Nikolai Dedok, whose location was allegedly traced via metro camera feeds tied into Kipod, leading to his arrest.

Results

Belarusian media, as cited by Privacy International, reported that more than 2,400 protest-linked criminal cases were opened in the year following the demonstrations. The European Union sanctioned Synesis and its CEO, Alexander Shatrov, in December 2020 for enabling security-force repression of peaceful demonstrators, with further sanctions reported in mid-2021.

Conclusions

This is documented here as a human-rights-harm case, not a success story: the system's facial-recognition capability was used to identify and prosecute peaceful protesters, prompting EU sanctions against its vendor. Synesis has disputed the characterisation of its role and contested the sanctions in court. A widely repeated claim about the total number of cameras nationally could not be independently verified and has been excluded rather than reported unverified.

Implementation
Indicative cost
Medium (€50k–€500k) — No budget figures disclosed; conservative estimate reflects a metro-and-public-camera-network-scale AI analytics deployment.
Time to results
Medium (1–3 years) — Documented harmful use is concentrated in the 2020-2021 period following the disputed August 2020 election; the underlying platform's broader operational history is not detailed in the sources reviewed.
Staffing & skills
Ministry of Internal Affairs of the Republic of Belarus, LLC Synesis (system developer/vendor)

Conditions for success

  • Integration of AI facial-recognition analytics with existing public and metro camera networks
  • Access to state camera infrastructure without documented independent legal or judicial oversight

Common failure modes

  • No public disclosure of the system's legal basis, data-retention rules, or independent oversight has been documented
  • The EU sanctioned the vendor (Synesis) and its CEO in December 2020, with further sanctions in mid-2021, for enabling repression of peaceful demonstrators
  • The system was used to identify and track a named activist (Nikolai Dedok), leading to his arrest, illustrating a direct human-rights harm rather than a public-value outcome

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful