Public bodies face a distinct set of rules about software that decides. Canada’s Treasury Board Directive on Automated Decision-Making requires impact assessments, notice, explanation and human intervention at levels set by the decision’s impact. Ontario’s Responsible Use of AI Directive and its Enhancing Digital Security and Trust Act, British Columbia’s FIPPA with the OIPC’s generative AI principles, Alberta’s new Protection of Privacy Act, Québec’s statement of principles and the OMB memoranda on federal AI use and acquisition in the United States all add specific duties. Texas’s Responsible AI Governance Act binds state agencies as well as businesses.
Conditionary’s government pack groups those instruments and asks a public administrator’s questions: which programs, which eligibility decisions, which enforcement actions, which public-facing assistants. It is honest about what it does not yet hold; the Ontario FIPPA rows and defence-sector content are on the watchlist, not quietly assumed.
For a municipality, the modelled Springfield, Texas workspace shows the scale: a city of 25,000 with a handful of decisions, a few systems, and a clean record it can show council.
Decisions the set-up guide asks about
- Benefits and eligibility determinations
- Permit and licence processing
- Enforcement prioritisation
- Public service chatbots
- Staff tools and knowledge work
Rules in the pack, among others
- Treasury Board Directive on Automated Decision-Making
- Ontario Responsible Use of AI Directive, EDSTA
- BC FIPPA and OIPC principles
- Alberta Protection of Privacy Act
- Québec Arrêté 2025-02
- OMB M-25-21 and M-25-22
- Texas HB 149
- Canadian and provincial human rights codes
Ready to run: Access and configuration changes gives an auditor general the who-changed-what record without a records request. See the reports.
