- Innovation
Girona automates the anonymization of personal data in public documents using AI: DigiCanvis
The Barcelona Provincial Council organized the “Conference on artificial intelligence in the local world. Local.iA: deployment of use cases” to showcase AI experiences applied to the field of local digital services.
During the conference, the AOC presented examples of use cases for applying AI in the field of algorithmic transparency, automatic categorization of telematic requests and the anonymization of documents.
The Head of Innovation Service at the AOC David Pastor, Dolores Caceres, from the Barcelona Provincial Council, explained what algorithmic transparency is and why it is key to generating trust in public artificial intelligence (AI). David Pastor presented the AOC's algorithmic transparency sheets, which explain in an understandable way how AI systems work, what data they use and what guarantees they offer. They also explained the risk management method applied before putting each AI system into service, to protect people, rights and the future.
Regarding the automatic categorization of telematic requests, the head of the Open Government service of the AOC, Ferran Farriol i Montserrat Fernandez, from the Barcelona Provincial Council, presented the two proofs of concept (POC) that are being worked on in the e-TRAM processing solution both in the citizen environment (procedure recommendation) and municipal internal (instance classification).
Through the eTRAM instance classification POC, the aim is to provide a classification proposal at the Classification Table (QdC) level based on the analysis of the content of the brief description of the instances. With the recommender POC, citizens are provided with a specific alternative within the generic instance catalog when they are carrying out the procedure. The AOC will continue to evolve the two POCs over the coming weeks to see if new modules can become available within the e-TRAM service.
In her speech, the head of Innovation service at the AOC Helena Carbonell, addressed a use case for improving internal efficiency: the document anonymization project. Helena Carbonell explained the privacy challenge posed by the automatic publication of plenary sessions on the transparency portal, and how the AOC is carrying out a proof of concept, together with 15 town halls, to test an automatic anonymization tool with artificial intelligence. She presented the key aspects of the tool, its operation and the future challenges that have been detected.