The next speaker in this session at the 5th Data Donation Symposium at the Weizenbaum-Institut in Berlin is Zoltán Kmetty, whose focus is on difference between participant intentions and actual donations to data donation projects. There is a considerable willingness-to-donation gap, and methodological interventions are required to address this; the present project tested this during the 2026 Hungarian parliamentary election by using AI chatbots as data donation assistants.
The study employed a standard data donation approach, but in the data donation module briefing split the participant base into two groups: one received a static data export guide, and the other an AI-assisted data export guide. As usual, it faced a substantial dropout of participants from those who accessed the donation site, clicked through the participant survey, and finally donated data and consented to their participation.
Users receiving AI assistance were slightly more likely to complete the data donation process than those receiving only static information; they also perceived the burden of providing data donations to be somewhat lower. These patterns did not depend on participants technological knowledge or attitudes towards AI. Interestingly, AI-assisted donations took some two minutes longer to complete than ordinary donations; this might have been due to the more interactive nature of AI chatbot engagement. Misuse of chatbot functionality was rare, and conversely it appeared to help especially where users encountered authentication errors during the data donation process.
