See something we could improve on this page? Give website feedback.
Label:
Hypothetical system — concept demonstration
A system record for a hypothetical AI system used in land-use review. This page is part of a concept demonstration and does not describe a system the City of Portland operates.
Status: Hypothetical system — concept demonstration. This system record is a labeled hypothetical addition to the City’s AI inventory, created for a Portland civic hackathon in August 2026. No such system exists.
| Operator | Portland Permitting & Development |
|---|---|
| Version | 0.3 (pilot) |
| First deployed | June 2026 (hypothetical) |
| Risk classification | HIGH — City AI Risk Model (BTS-4.04 Guidance, updated Feb 2026) |
Purpose
Screens land-use applications and city records to recommend review priority and possible conditions. It does not decide.
Risk classification
This system is classified HIGH under the City AI Risk Model (BTS-4.04 Guidance, updated Feb 2026). High-risk systems are:
"Subject to the highest degree of review, evaluation, and controls by BTS. Implementors must comply with requirements for testing, training data validation, cybersecurity safeguards, ongoing monitoring, and audits."
What this system may and may not do
| Allowed ✓ | Not allowed ✗ |
|---|---|
| Read application materials and city records | Approve or deny applications |
| Flag applications | Alter records |
| Recommend conditions | Contact applicants |
| — | Issue decisions |
BTS-4.04 prohibits automated decision-making "without an appropriate level of human review, proportionate to the risk" where outcomes could significantly affect "an individual's health, safety, rights, or financial well-being."
Information used
| Information used | Source |
|---|---|
| Application materials | Submitted plans |
| Property records, including year built | County Assessor import |
| Zoning and overlays | City GIS |
| Prior permits | PP&D records |
Not used: applicant identity or demographics; property value; neighborhood income.
Human review
Every recommendation is reviewed by City staff. Final authority rests with the responsible bureau official.
Aggregate record
Fictional figures, shown to illustrate what a public system record would report.
| Applications screened | 1,284 |
|---|---|
| Applications flagged | 173 |
| Recommendations overridden by staff | 38 |
| Data corrections reported by residents | 12 |
| Outcomes changed after correction | 3 |
Decisions involving this system
Report a concern about this system
To report a concern, a data error, or a question about this system, see AI at the City.