May 10, 2026 · 4 min read

How AI Is Changing How People Choose Who to Vote For

AI voter guides are offering a better alternative to endorsement lists and party cues. Here's what they get right, what they get wrong, and how they work.

For most of American electoral history, deciding who to vote for in down-ballot races meant one of three things: following your party's endorsements, looking up your local newspaper's picks, or guessing in the voting booth. All three approaches have significant problems. AI is starting to offer a fourth option — and it's meaningfully better.

The down-ballot problem

Presidential and gubernatorial races get massive media coverage. Voters hear about candidates for months before election day. But the further down the ballot you go, the less information reaches the average voter.

In a typical LA primary, fewer than 20% of voters have any meaningful knowledge of the City Attorney candidates before they vote. For County Assessor, the number is lower. For school board races, many voters skip the line entirely rather than vote uninformed. This isn't a failure of civic engagement — it's a failure of information infrastructure.

Party endorsements

Assumes you agree with everything your party does

Newspaper picks

Editorial board has its own institutional perspective

Guessing

The actual outcome for most down-ballot races

What AI actually does differently

The new generation of AI voter tools — of which WhoDoIVoteFor is one — takes a different approach. Instead of presenting you with a list of candidates and leaving you to research each one, they start with you.

What do you care about? Homelessness, housing, crime, schools, taxes, climate, city services? How do you think about those issues specifically — not in the abstract, but in the concrete choices candidates have to make? Once you've articulated your concerns, the AI does the research work: matching each candidate's stated positions, voting record, and endorsements against your specific priorities.

The output isn't a generic endorsement — it's a personalized alignment score that shows you exactly where each candidate matches your values and where they don't.

The nonpartisan advantage

Traditional voter guides have an inherent bias problem: they're produced by organizations with their own political perspectives. The League of Women Voters skews one way. The Chamber of Commerce skews another. Newspaper editorial boards have their own institutional perspectives.

An AI voter guide trained on candidate positions across the full political spectrum — covering both progressive and conservative concerns with equal rigor — can present alignment data without an editorial thumb on the scale. If you care about encampment enforcement, the AI tells you which candidate is most aligned with that position. If you care about tenant protections, same thing. The system doesn't have an opinion about which concern is more legitimate.

The limits of AI in civic decision-making

Honest disclosure: AI voter guides have real limitations.

⚠️

Data quality matters. Candidate data is only as good as the research behind it. A well-funded candidate with detailed policy positions will be matched more accurately than a school board candidate who has said little publicly.

⚠️

AI captures positions, not character. A candidate who says the right things might govern very differently from their platform. An AI can't tell you that.

⚠️

Voting is still a values judgment. The AI can tell you which candidate's positions are most aligned with your stated concerns — it can't tell you which of your concerns should matter most.

What it gets right

What AI does well is something no traditional voter guide can: it processes the full matrix of candidate positions across every race simultaneously, weights them against your specific priority set, and surfaces alignment you wouldn't have found by reading endorsement lists. For the races you'd otherwise skip — City Attorney, Assessor, School Board, the five ballot measures — it gives you a starting point grounded in something you actually care about, rather than a coin flip or a party line. That's not a revolution in democracy. But it's genuinely better than guessing.

Try it for the June 2 primary

WhoDoIVoteFor covers all 13 races on the LA ballot — 68 candidates, hand-researched, matched by AI against your concerns. The Mayor's race is free. Everything else is $4.99. Takes two minutes.

Ready to vote?

Find your candidate match in 2 minutes

Enter your ZIP, pick your concerns, get AI-matched results for all 13 races on your ballot.

Find your candidate match →