The Race — Building Democracy in the Age of AI. People race toward a shared finish line.
The Race
← Trends

Established trend · Active

AI screening becomes an institutional gateway before human decisions

Latest assessment: · Evidence updated

Standing score
Unscored
Direction of power
Mixed-contested
Window
Unclear
Confidence
Moderate

What we think is happening

Permitting AI increasingly sits before the official reviewer: applicants are asked, contracted or required to pass through an automated completeness/compliance check. Final legal authority can remain public while the practical entry process becomes vendor-mediated.

Scope: Municipal residential permitting; several buyers of a correlated vendor.

Potential direction of power

Honolulu’s requirement, Denver’s contract and Seattle’s pilot show different stages of AI screening becoming part of public services. Such systems can ease navigation while creating new dependencies. Service improvements, routes to challenge a result, alternatives and retained city expertise remain unevenly demonstrated.

Supporting cases

Honolulu DPP

Honolulu made CivCheck pre-application review mandatory for eligible residential permits from September 1. Official staff retain final permit authority; mandatory screening creates a new practical gateway.

Read the case and sources →

Seattle Department

Seattle compared CivCheck screening against staff review in a 2025 pilot and reported results June 17, 2026. Checks reached 87% completeness accuracy and 92% compliance accuracy, but incomplete coverage could leave the same number of review cycles.

Read the case and sources →

Denver,

Denver approved about $4.6 million over five years for CivCheck intake/plan-review support on March 9. Contract approval is verified reporting; the record’s operational/scaling description does not establish complete rollout or measured outcomes.

Read the case and sources →

A procurement case shows why AI-generated evaluations belong in the review record

In the TRAX procurement challenge, the Court of Federal Claims required three AI-generated proposal evaluations to be included in the record on July 10. Its later merits opinion, made public September 22, described their circulation to people involved in the Army's evaluation of a roughly $449 million support contract. An October 6 legal explanation brought the case back into view.

TRAX ultimately lost its challenge. The court's treatment of the documents did not establish that AI evaluation is always lawful or unlawful; it showed that the government's description of how material was used can be tested against the actual record. The distinction between generating advice, sharing it and relying on it mattered.

For The Race, this is a concrete accountability mechanism inside an existing institution: procurement review can expose AI's role rather than accepting a claim that it was irrelevant.

Distinct connection: AI-generated evaluations enter the human evaluation/decision gateway. Overlapping evidence is not independent confirmation.

Read the case and sources →

Window of contestability

Unclear for applicants and cities: pre-screening may ease entry but mandatory vendor steps may add a gate; human authority and alternative suppliers are relevant counterweights.

What would change this assessment

Clear human override and complaint routes, usable alternatives, practical handover terms and independently measured service outcomes would clarify who controls the gateway.

Related cases

← All 24 trends