Cities Turn to AI for Faster Permits, But CDT Warns of Due Process Risks
Local governments across the US are using AI to cut building permit wait times, and federal money is now backing the approach. The Center for Democracy & Technology (CDT) argues that the speed gains come with risks to due process, fairness and vendor accountability.
- HUD issued AI permitting recommendations in May 2026 and offered up to $3 million in grants a month later.
- Honolulu's median building permit wait fell 40%, from four months to two and a half months, according to CDT's analysis of city data.
- Hernando County, FL cut post-Helene permit processing from 30 days to two using AI.
- CDT warns of opaque denials, algorithmic bias and weak vendor oversight.
- CDT urges human oversight, public transparency and a right to contest decisions.
Federal support and early deployments
According to CDT, a shortage of affordable housing is linked to a failure to build enough homes, with burdensome permitting and regulatory friction often cited as causes. In 2025, AI emerged as a possible fix for long waits on environmental and building permits in municipalities such as Austin, TX, Honolulu, HI, and Los Angeles, CA.
In May 2026, the Department of Housing and Urban Development (HUD) released recommendations for state and local practices that encouraged municipalities to use AI to streamline construction permitting timelines. A month later, HUD offered up to $3 million in grants for local governments to deploy automated permitting and building code systems. The funding covers the first three years of software licensing plus staffing costs to run the tool, using platforms such as PermitFlow, Blitz Permits, CivCheck, or Permitify.
Three ways cities use AI
CDT describes three main functions, noting that federal funding and local pilots are still at an early stage.
- Error detection: Denver, CO, Jacksonville, FL, and Port Orchard, WA use AI as a pre-screening filter that catches application mistakes and cross-checks building codes before formal submission.
- Automated document analysis: Seattle uses a "copilot" model, where AI assists a human reviewer rather than replacing them or making the final decision. It uses computer vision to scan plans and intake documents for completeness.
- Accelerated plan review: Los Angeles adopted software that can evaluate complex building plans and generate zoning compliance reports in about a single business day.
CDT points to early results. In Honolulu, after a rocky start, data from the Department of Planning and Permitting showed median wait times for building permits fell 40%, from four months to two and a half months. Lancaster and San Jose, both in California, procured AI tools that aim to cut review times for simpler projects from months to days.
The tools may also help planning departments that CDT says face a 15-year staffing crisis driven by a shrinking workforce and more complex regulation. After Hurricane Helene, Hernando County, FL used AI-driven technology to reduce post-disaster permit processing times from 30 days to two. Los Angeles deployed similar tools after the wildfires.
The risks CDT highlights
CDT warns that automated decision-making in public services still threatens basic due process and transparency. Algorithmic denials of permits or public benefits, it says, often leave people without written explanations or meaningful ways to appeal. It describes a shift toward "black box" systems whose inner mechanics stay hidden.
CDT also cites HUD's own observation that local agencies sometimes adopt these tools despite a lack of "publicly available documentation of implementation experiences, staffing implications, and governance considerations."
On bias, CDT argues that replacing human discretion with an unscrutinized algorithm does not remove bias. Without safeguards such as human rights impact assessments and public grievance channels, it says, automated systems tend to reproduce and amplify historical biases, which can be harder to combat than human ones. Weak vendor management adds to the problem, since municipal contracts routinely omit basic standards for algorithmic fairness.
What CDT recommends
CDT says responsible use requires more governance. It cites Honolulu's planning department, which intentionally avoids fully automated reviews and single-click algorithmic decisions. To scale such models, CDT says state and local agencies should build frameworks that bring civil rights, legal and risk officers into the AI deployment lifecycle.
The organization also calls for public transparency about algorithmic use and meaningful human oversight that preserves a person's right to contest decisions. It adds that sustainable funding and long-term oversight beyond temporary grant windows, such as HUD's three-year clock, will help agencies modernize without compromising due process.
For cities weighing these tools, the message is that speed is only part of the picture: CDT concludes that robust safeguards must be established before implementation.
Center for Democracy & Technology: Faster Permits, Fewer Eyes: The Case for and Against AI Review