Transparent and Trustworthy Digital Democracy

Democratic participation increasingly depends on citizens’ ability to interpret and evaluate complex political communication and decision mechanisms. Political discussion has become increasingly polarized, fragmented, and vulnerable to noise and manipulation, resulting in people distrusting the democratic system and its actors and disengaging from civic participation.

Beyond political discourse, democratic systems face structural computational challenges that current civic technologies do not address. Democratic procedures must aggregate preferences, allocate resources, and ensure equitable representation.

To address these challenges, this project works on:

• Leveraging AI to interpret political text into ideological structure and provide users with evidence-grounded explanations of how candidates and institutions position themselves across policy domains.
• Developing methods for analyzing uncertainty in voting systems when ballots are delayed, missing, or only partially observed. This work provides more transparent reporting, auditing, and explanation of ranked-choice election outcomes, particularly in settings where partial results and delayed ballots have contributed to public confusion and mistrust.