In January, the Ontario Human Rights Commission and the Ontario Privacy Commissioner jointly released six principles for the responsible use of AI.
The combination is unusual. Most AI guidance is written by one type of regulator; privacy, safety, sectoral. Having a privacy body and a human rights body co-sign the same document is rare, and it broadens what gets covered.
The six principles:
- Valid and reliable: does the AI actually do what it's said to do, consistently?
- Safe: does its use protect people from foreseeable harm?
- Privacy protective: is personal information collected, used, and disclosed lawfully?
- Human rights affirming: does it respect equality, dignity, and freedom from discrimination?
- Transparent: can affected people understand that AI is being used, and how?
- Accountable: is there a clearly identified human or organization on the hook?
None of these are new ideas. What's useful is that they're now in one place, signed by two Ontario authorities, and short enough to print.
If your organization is deploying AI in any capacity that touches Ontarians, this is a reasonable place to start a self-assessment.
What this means for health organizations
The six principles work as a short self-assessment for each AI tool you use: one line per principle, with the evidence you'd point to. The gaps show where to focus first.
