Qualitative analysis of AI data-centre backlash

Centralize hearings, petitions, transcripts, and public records to produce defensible, segment-aware insights on AI data-centre opposition within a rapid pilot workflow.

Automated transcription, AI-assisted coding, and cross-segment frequency analysis let research and policy teams quantify local concerns and produce stakeholder-ready briefs quickly.

• CBC reported on June 7, 2026 that Canada has five existing hyperscale AI data centres and 96 proposed or under construction projects.
• A United Nations report estimates data centres consume about 448 TWh of electricity, produce roughly 189 million tonnes of CO2, and use about 4.5 trillion litres of water, figures that fuel local opposition.
• A recent poll (Angus Reid, reported by CBC) found 68% of Canadians oppose a large AI data centre near where they live.
• A Manitoba petition gathered 13, 500 signatures, illustrating how local mobilization can influence project outcomes.

Proposals for large AI data centres across Canada triggered protests and regulatory pushback in early to mid 2026.

Across Canada, proposals for large AI data centres triggered protests and regulatory pushback in early to mid 2026. Notable events included a Hamilton planning hearing with hundreds attending that resulted in a denied application, marches in Vancouver focused on water and energy use, a Regina demonstration at the legislature, and a Manitoba petition that led provincial leaders to halt a project.

• Primary concerns driving opposition: land use, electricity demand, water for cooling, health and perception issues, and broader public anxiety about jobs and culture.
• Public numbers cited by researchers and UN reports make the environmental case tangible, energy and water metrics are headline drivers of local mobilization.

Run this two-week pilot to produce a stakeholder-ready brief.

• 1) Gather inputs: council minutes, public submissions, local news (for example, CBC articles), protest photos, petition data.
• 2) Import to Evidano and run automated transcription and cleanup; apply a custom dictionary for local terms.
• 3) Auto-generate initial themes; review and refine a lightweight codebook with domain experts.
• 4) Run cross-segment frequency analysis (by municipality, distance to site, demographic group).
• 5) Pull representative verbatim quotes and sentiment co-occurrence networks for each theme.
• 6) Validate with 5–10 targeted AI-avatar interviews or manual follow-ups where gaps remain.
• 7) Produce a two-page decision memo: top concerns, quantified evidence, recommended mitigation experiments.

Centralize your corpus and run an initial thematic and cross-segment analysis to move beyond media narratives and produce defensible insights.

• Ready to try a pilot? Try Evidano for free to import your first case files and run a rapid two-week analysis that delivers theme frequencies, quotes, and stakeholder-ready visuals.
• For research ethics: treat this as non-diagnostic, consent-sensitive work and follow local privacy rules when handling personal submissions.