Incident 961: Les autorités serbes auraient utilisé des outils Cellebrite basés sur l'IA pour déverrouiller le téléphone d'un journaliste et installer un logiciel espion.
Description: Les autorités serbes auraient utilisé les outils d'analyse forensique de Cellebrite, basés sur l'IA, pour déverrouiller les téléphones de journalistes et de militants sans leur consentement. Elles auraient ensuite installé NoviSpy, un logiciel espion récemment découvert. Ce logiciel aurait ensuite permis l'extraction secrète de données, l'activation à distance du microphone et de la caméra, et la surveillance. Amnesty International a découvert des preuves forensiques reliant l'Agence serbe d'information sur la sécurité (BIA) à ces attaques. Cellebrite a suspendu ses ventes à la Serbie après la publication de ce rapport.
Editor Notes: Amnesty International's report can be accessed at the following URL: https://www.amnesty.org/en/documents/eur70/8813/2024/en/. The BIA's brief response to the report can be read here: https://www.bia.gov.rs/mediji/saopstenja-za-javnost/saopstenje-bia-16-12-2024-godine/. In terms of the timeline of events, it is reported Serbian authorities were using Cellebrite between 2021 and 2024. In February 2024, they reportedly targeted the journalist Slaviša Milanov; between July and November 2024 there were reportedly at least nine activists who were detained or questioned and had their phones and computers seized, with allegations that the Serbian security services used the Cellebrite UFED forensic tools to extract data from the devices and, reportedly in some cases, they had installed NoviSpy. Between July and August 2024, when anti-lithium mining protests took place, Serbian authorities reportedly arrested 33 activists and seized their devices for forensic searches. In October 2024, an activist from the NGO Krokodil had, according to Amnesty International's forensic analysis, NoviSpy installed on their device while being interviewed by the BIA. On December 16, 2024, Amnesty published their report (which is being used for this incident ID's date). In response, Cellebrite halted their sales to Serbia, which they announced on February 25, 2025. It is also reported that Google Project Zero and the Threat Analysis Group have confirmed a zero-day vulnerability in Android exploited by Serbian authorities, which was later patched.
Outils
Nouveau rapportNouvelle RéponseDécouvrirVoir l'historique
Le Moniteur des incidents et risques liés à l'IA de l'OCDE (AIM) collecte et classe automatiquement les incidents et risques liés à l'IA en temps réel à partir de sources d'information réputées dans le monde entier.
Entités
Voir toutes les entitésAlleged: Serbian Security Information Agency (BIA) et Cellebrite developed an AI system deployed by Serbian Security Information Agency (BIA) , Serbian police et Government of Serbia, which harmed Slaviša Milanov , Nikola Ristić , Krokodil , Journalists in Serbia , Human rights defenders in Serbia , Environmental activists in Serbia , Dissidents in Serbia et Civil society organizations in Serbia.
Statistiques d'incidents
Risk Subdomain
A further 23 subdomains create an accessible and understandable classification of hazards and harms associated with AI
4.1. Disinformation, surveillance, and influence at scale
Risk Domain
The Domain Taxonomy of AI Risks classifies risks into seven AI risk domains: (1) Discrimination & toxicity, (2) Privacy & security, (3) Misinformation, (4) Malicious actors & misuse, (5) Human-computer interaction, (6) Socioeconomic & environmental harms, and (7) AI system safety, failures & limitations.
- Malicious Actors & Misuse
Entity
Which, if any, entity is presented as the main cause of the risk
AI
Timing
The stage in the AI lifecycle at which the risk is presented as occurring
Post-deployment
Intent
Whether the risk is presented as occurring as an expected or unexpected outcome from pursuing a goal
Intentional
Rapports d'incidents
Chronologie du rapport
Loading...
