Policies Public-facing policies
AI Output Accuracy Policy
Standards Map.ca holds AI-generated and AI-assisted outputs to before they reach users.
- Version
- 0.1.0
- Effective
- May 20, 2026
- Last reviewed
- May 20, 2026
- Review cycle
- every 6 months
- Master Policy Index entry
- §4 #32
Policy text
AI is fast and confident. That is exactly why Map.ca cannot ship AI outputs without an accuracy standard. The AI Output Accuracy Policy defines the categories of AI output (drafts, suggestions, summaries, translations, classifications), the required accuracy posture for each, and the feedback path users have when an AI output is wrong.
It applies to every Map.ca surface that uses AI to produce user-visible content, and to the AI team that tunes those models.
Principles this policy enforces
- AI assists, people remain responsible
- Consent must be meaningful
- Public data and personal data are not the same thing
What it requires
- Distinguish AI drafts from AI conclusions in user-visible surfaces.
- Give users a clear path to report AI errors.
What it forbids
- Do not present AI suggestions as verified facts.
- Do not suppress AI error reports to protect performance metrics.
How it applies
- Map.ca AI team
- Product teams using AI outputs
- Users encountering AI outputs
References
- Government of Canada — Voluntary Code on Generative AI (Validity and Robustness)
Related policies
AI Use Policy
Where AI is used on Map.ca, what it can and cannot do, and how humans stay responsible for outcomes.
AI Transparency Label Policy
Where and how AI usage is disclosed to users across Map.ca surfaces, plus the hybrid hosting and data-residency model behind it.
AI Human Oversight Policy
Which AI-involved decisions must route through a human reviewer before becoming final.