How AI can save democracy -- not destroy it

Can Your Ballot Withstand the Algorithm?

What if the next election you participate in is already being shaped by systems you can't see—systems that might amplify doubt or clarify truth? For anyone who has scrolled through targeted political posts and wondered whose voice is truly behind them, the question of what specific AI tools could protect election integrity against manipulation feels urgent and personal.

Divya Siddarth confronts this exact tension in her talk. She describes AI's dual role in elections, moving from the well-known risks of misinformation to the potential of carefully built tools that reinforce democratic participation. Her core claim is that thoughtfully designed AI systems can strengthen transparency, participation, and trust rather than undermine them. This argument rests on the premise that technology itself carries no fixed destiny; its impact depends on deliberate choices about how it is applied to civic processes.

Siddarth's position directly challenges the narrative of a silent coup in which AI quietly distorts electoral outcomes. Instead of accepting that reshaping as inevitable erosion, she reframes AI as a possible corrective force. Where fears center on manipulation and eroded confidence, her reasoning highlights pathways toward greater openness and citizen involvement, turning the same technology that spreads doubt into one that could verify information and broaden engagement.

• The speaker emphasizes design intent as the deciding factor between harm and benefit.
• She positions participation and trust as measurable outcomes that AI can support when aligned with democratic goals.
• Her view rejects technological determinism, insisting human decisions about deployment matter most.

Yet this optimistic reframing leaves a lingering tension: even if the tools exist in principle, who will ensure they are built and governed to serve the public rather than narrow interests? Siddarth's perspective invites ongoing scrutiny of whether societies will prioritize those protective designs before manipulation outpaces them.