By the year 2045, demographers predict there will be no majority ethnic group in the United States, and artificial intelligence (“AI”) technologies will likely shape the trajectory of multiracial democracy. While AI has the potential to meaningfully support our multiracial democracy (reduce costs and allow for expanded democratic participation, more effective content moderation, cross-racial deliberation and coalition building, more equitable policymaking, and fairer electoral districts and voting rules), AI can also fuel targeted disinformation campaigns, racial polarization, voter suppression, discriminatory gerrymandering, and cyberattacks. Even absent intentional discrimination, bias embedded into AI and data used in voter registration list maintenance, signature matching used to verify mail-in-ballots, picking polling locations, designing ballots, content moderation, detecting deepfakes, large language models, and other tools of the election process could replicate racial and ethnic hierarchy in elections and policymaking well into the future. Unfortunately, competitive advantage and financial incentives prevent private sector companies and elected officials from consistently acting to prevent these harms, and sometimes drive them to exacerbate the harms. This session sketches out the essential role of philanthropy in ensuring that AI is developed, applied, and regulated to facilitate a well-functioning multiracial democracy.

​