Discussion

Voters are asking AI chatbots who to vote for. The answers can miss the mark

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#8890

Voters are turning to AI chatbots for election information, but the answers can leave out candidates or rely on shaky material. A Canadian Press report describes one Toronto voter using Claude to compare local candidates, alongside researchers’ warning that chatbots have already given voters false or outdated information in Quebec.

Watch Desk analysis

What happened

Eric Apps, a Toronto lawyer, asked Claude for the five leading candidates in his ward and their positions on bike lanes, housing and property taxes. He found the answer easy to digest, but said the chatbot had little to offer on candidates with less media coverage or a smaller online presence.

The Canadian Press reports that McGill University researchers estimated 40 per cent of Quebec voters turned to AI chatbots during the province’s recent election and received false or outdated information. The report does not give further detail on the researchers’ method or what counted as an incorrect answer. Western University professor Alissa Centivany also warned that chatbot responses can draw on social-media content and reflect a user’s previous queries. Anthropic said Claude uses web search, shows its sources and is trained to flag disputed claims or conflicting sources.

Why it matters

Local elections can be especially hard to research: in Ontario municipal races, candidates do not run under party labels, so voters may have to compare unfamiliar names and policies themselves. A chatbot can make that task quicker, but a neat summary is not the same as a complete or reliable one. Candidates who leave little online footprint may be the very ones an answer fails to represent.

Our read

Chatbots may help people get oriented, and the Toronto voter’s experience shows why they are appealing. But use them as a starting point, not as the ballot-box oracle. Check candidate claims against official election information and the candidates’ own material, particularly when an answer presents a confident shortlist without explaining who is missing.

What to watch

  • Whether researchers publish more detail on how they measured chatbot use and incorrect election answers in Quebec.
  • How chatbots handle lesser-known candidates and cite the material behind their summaries.
  • Whether voters treat AI answers as a first stop or a substitute for checking candidates directly.

Discussion spark: Should election chatbots be expected to surface lesser-known candidates fairly, or is it the voter’s responsibility to check who the summary leaves out?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.