Most artificial intelligence (AI) governance guidance in Canada focuses on high-level, principle-based approaches, emphasizing “responsible,” “safe” and “ethical” AI. However, these approaches leave governance gaps in many sectors, including housing. As the second part of a two-part working paper series, this paper offers five policy recommendations to close this gap and mitigate AI-related harm in housing:
- Establish a justification requirement for federal entities (and others that benefit from federal funding for housing) that are looking to use AI in the context of shelter service and housing provision.
- Mandate human rights impact assessments as a condition of government funding for organizations that produce or use AI in homelessness management, either as a product or in operations.
- Do not mandate the use of AI in homelessness management as a condition for government funding.
- Prohibit the use of facial recognition technologies in residential areas.
- Require both in-person and online hearings be made available for landlord-tenant dispute hearings.
The paper then ties these AI- and data-focused recommendations to two broader housing policy suggestions: the promotion of community land and data trusts and community-led response.
