AI Chatbots Become New Voter Resource Amid Misinformation Risks
AI-generated image · Bay Street Wire
As Toronto voters navigate a crowded municipal field, experts warn that reliance on AI for candidate research may prioritize user preference over objective facts.
Voters in Toronto's 2026 municipal election are increasingly turning to artificial intelligence to navigate a crowded political landscape. In the Parkdale-High Park ward, where longtime councillor Gord Perks retired and the field grew to 10 candidates, lawyer Eric Apps used Anthropic’s Claude chatbot to summarize candidate policies on housing, property taxes, and bike lanes, according to Global News Toronto.
While Apps found the tool provided a digestible format for political participation, researchers warn of systemic risks. Alissa Centivany, an associate professor at Western University, told Global News Toronto that chatbots are replacing traditional search engines and reliable media. Centivany noted that AI may compile information based on what it believes a user wants to hear rather than objective facts, often incorporating inflammatory or untrue user-generated content from sites like Reddit.
These concerns follow a provincial election in Quebec where McGill University researchers estimated 40% of voters received false or outdated information from AI chatbots. The inaccuracy is compounded by the fact that many Quebec news outlets block chatbots from scraping their websites, Global News Toronto reports.
Anthropic has since updated its usage policies to prohibit using Claude to deceive voters, impersonate officials, or spread false information to suppress turnout. The company stated that Claude is designed to flag disputed claims and show sourcing so users can judge credibility. Alex Marland, a professor at Acadia University, told Global News Toronto that voters are drawn to these tools because Ontario municipal candidates are not affiliated with parties, requiring voters to invest more individual effort to understand their choices.