Comune di Trento
Marvel and Protector AI surveillance research projects
Italian regulator fined the city of Trento EUR 50,000 over two EU-funded AI surveillance projects using street microphones and social media
Problem
The city took part in EU-funded smart-city research on detecting public-safety risks from cameras, microphones and online posts.
Approach
Marvel combined existing city CCTV with audio from street microphones, and AI analysed it to detect risk events. Protector used CCTV without sound plus hateful social media posts to flag threats to places of worship. The city said data were anonymised right after collection. The Garante decided on 11 January 2024 that the city had no legal basis, that anonymisation was insufficient and that notices did not say microphones could record conversations. It also found no impact assessment, fined the city and ordered deletion of the data.
Results
Takeaway
Taking part in EU research did not give the city a legal basis to process residents' data. The regulator said it was open to talks on AI projects that follow privacy rules.
Categories
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer visionPredictive analytics and ML
- Adoption
- Government support program
- Goal
- Quality and risk
- Evidence
- Regulator, audit or court
Sources
Source summaries
The authority's release on its 11 January 2024 decision. It fined Trento EUR 50,000 and ordered deletion of unlawfully processed data. Grounds include no legal basis, data shared with partners, weak anonymisation, unclear notices about microphones and no proven impact assessment.
Local newspaper report on the fine. It describes Marvel (CCTV plus street microphones) and Protector (CCTV plus social media hate messages), notes the city's claim of immediate anonymisation and the Garante's rejection of it.
Full details are in the original sources.