How to use AI in analysing mass responses to FOI requests

Longer description:

The issue with public debates is that they often rely on unverifiable or outdated data. Public administration bodies hold a wealth of information that could clarify societal discussions if properly utilised. However, the formats this information is held in often makes it challenging to interpret digitally.

The civic data mining project Fedrowanie aims to streamline the gathering of information from public institutions through mass FOI requests. It organises the collected information into ‘monitorings’, thematic groups of queries. Each monitoring involves several letters exchanged with offices, aiding thematic organisation and follow-up.

Fedrowanie was successful in collecting data, but it struggled to organise it for analysis. To address this, the CheckHowItIs system was developed, which enables volunteers to convert documents into a standardised, analysable format. This process requires volunteers to interpret responses and input data into a database, which is then used for research to enhance social awareness and governance quality.

Despite challenges like campaign completion and data inconsistencies, the initiative has fostered community building and directed efforts towards tangible goals. With advances in AI, the Fedrowanie project transitioned to an AI module to process documents more efficiently and accurately, marking a significant improvement in public data analysis and governance.This evolution from manual volunteer work to AI integration in the Fedrowanie project represents a leap in enhancing public discourse and governance.