AI-Assisted Permitting and Application Review Act

Task Forces: Commerce, Insurance and Economic Development

Draft Disclaimer

Prior to task force meetings, ALEC posts these legislative member-submitted draft model policies to our website. The draft model policies are then discussed, debated, and voted on by ALEC task force members. Policies that receive final approval by legislators on the ALEC Board of Directors become official ALEC model policy. Draft model policies that fail to become official ALEC model policy are removed from the website.

Summary

The AI-Assisted Permitting and Application Review Act lets state agencies use artificial intelligence tools to speed up the review of permit, license and regulatory applications. Faster, more predictable approvals lower the cost of starting and growing a business, attract investment and help agencies handle rising workloads without proportional staff growth. The Act creates a pilot program for agencies. It sets minimum standards for data security, U.S. data residency, vendor experience and platform design, and exempts pilot contracts from standard state purchasing laws so agencies can move quickly. It also requires public reporting on processing times, accuracy and costs. AI tools may check applications for completeness, flag compliance issues and prepare draft review summaries. Every permit decision stays with a qualified human reviewer, and AI may never approve, deny or condition a permit on its own.

Model Policy

Section 1. Title.

This Act may be cited as the “AI-Assisted Permitting and Application Review Act.”

Section 2. Legislative Findings and Purpose.

(A) The Legislature finds that:

  1. Delays and unpredictability in state permitting and licensing raise costs for businesses, discourage investment and slow job creation.
  2. State agencies face growing application volumes, staffing shortages and complex regulatory requirements that lengthen review times.
  3. Many applications are returned because they are incomplete or contain correctable errors, adding rounds of review for both applicants and staff.
  4. Artificial intelligence tools can check applications for completeness and compliance, summarize lengthy submissions and help staff focus on substantive judgment.
  5. Public trust requires that AI tools be secure and transparent, and that permitting decisions remain with accountable public employees.

(B) The purpose of this Act is to improve the speed, consistency and accuracy of state permitting and licensing by authorizing the responsible use of AI-assisted review tools, while keeping human oversight and final decision authority with agency staff.

Section 3. Definitions. As used in this Act:

  1. “Agency” means [the Department of Environmental Quality, the Department of Commerce, the Department of Labor and Licensing, and any other executive agency designated by the Governor or the Chief Information Officer] that issues permits, licenses, registrations, certifications or plan approvals.
  2. “AI-assisted review platform” means software that uses artificial intelligence, including machine learning or large language models, to help agency staff or applicants evaluate applications against applicable statutes, rules and guidance.
  3. “Applicant-facing pre-submission tool” means a function that lets an applicant check a draft application for completeness and potential compliance issues before formally submitting it.
  4. “Staff-facing review tool” means a function that helps agency staff evaluate a submitted application, including completeness checks, issue identification, document summaries and draft review comments.
  5. “Application” means any request for a permit, license, registration, certification, variance or plan approval, including its narrative descriptions, forms, plans, technical reports, financial disclosures, certifications and supporting documents, whether typed or handwritten.
  6. “Final decision” means any approval, denial, issuance, modification, condition, suspension, revocation or determination of completeness that has legal effect on an applicant.
  7. “Qualified reviewer” means an agency employee or licensed professional authorized under state law or agency rule to make or recommend a final decision.
  8. “Department” means [the Department of Information Technology / Office of the Chief Information Officer].

Section 4. AI-Assisted Permitting Pilot Program Established.

(A)Within [XX] days after the effective date of this Act, [each agency / the agencies designated in Section 3], in consultation with the Department, shall establish a pilot program to use AI-assisted review platforms to modernize permit and license review, reduce processing times and manage increasing workloads.

(B) The pilot program shall include at least [two] permitting or licensing programs that, together, account for a significant share of the agency’s application volume or backlog.

(C) The pilot program shall run for at least [12] months. At its conclusion, the agency may continue, expand or end the use of AI-assisted review based on the results reported under Section 9.

(D) The agency shall provide training to qualified reviewers on the capabilities, limitations and proper use of the platform before staff use it on live applications.

Section 5. Authorized Uses.

(A) An agency may use an AI-assisted review platform to:

  1. Let applicants check draft applications for completeness and potential compliance issues before submission.
  2. Check submitted applications for completeness against statutory and regulatory requirements.
  3. Identify potential compliance issues, inconsistencies, missing information or applicable rules for staff attention.
  4. Summarize narrative applications, technical reports, financial disclosures, certifications, structured forms and handwritten materials.
  5. Prepare draft deficiency letters, review comments or decision documents for review and editing by a qualified reviewer.
  6. Route, prioritize and track applications, and report on processing times.

(B) An applicant’s use of a pre-submission tool shall be voluntary. A favorable or unfavorable result from a pre-submission tool shall not bind the agency and shall not constitute a final decision.

Section 6. Human Oversight and Final Decision Authority.

(A) An AI-assisted review platform shall not make a final decision. Every final decision shall be made by a qualified reviewer, who shall review the application and any AI-generated output on which the decision relies.

(B) A qualified reviewer may accept, modify or reject any AI-generated finding, summary or draft. The agency shall not adopt any performance standard, quota or practice that requires or pressures staff to accept AI-generated output without independent review.

(C) The agency record for each application shall identify the qualified reviewer responsible for the final decision.

(D) An application shall not be denied, deemed incomplete or returned solely on the basis of an AI-generated finding.

(E) Nothing in this Act changes the legal standards, public notice and comment requirements, or appeal rights that apply to any permit or license.

Section 7. Procurement and Vendor Qualifications.

(A) The agency shall contract with a third-party vendor to furnish an AI-assisted review platform.

(B) A vendor shall meet all of the following minimum qualifications:

  1. The vendor and any parent entity are incorporated and headquartered in the United States, and are not owned, controlled or directed by a foreign adversary as defined in [15 C.F.R. 791.4 / state law].
  2. The platform can provide AI-assisted review across multiple permitting or licensing programs, not only construction or building permits.
  3. The vendor has deployed or piloted an AI-assisted review platform with at least one state-level government agency.
  4. The platform provides both an applicant-facing pre-submission tool and a staff-facing review tool, with data shared between them so that review of a submitted application can build on its pre-submission check.
  5. The platform can review the full range of documents in an application, including narrative text, financial disclosures, compliance certifications, structured forms and handwritten materials, not only drawings or design files.
  6. The platform records the sources, rules and reasoning behind each AI-generated finding so a qualified reviewer can verify it.
  7. The platform features an AI-native architecture, purposefully designed to use AI technology. A legacy platform retrofitted to enable the use of AI technology does not meet this qualification.

(C) In evaluating proposals, the agency shall give significant weight to:

  1. Demonstrated data security and privacy protections.
  2. Documented improvements in accuracy and processing times for government customers.
  3. Total cost of ownership, including implementation, training and ongoing support
  4. Ease of integration with the agency’s existing permitting and records systems.

Section 8. Data Security and Privacy.

(A) All data collected, stored or processed under a contract entered into under this Act shall be stored on servers physically located in the United States.

(B) Application data remains the property of the state. The vendor shall not sell application data.

(C) The vendor shall notify the agency of any security breach involving state data within [72] hours of discovery.

(D) Confidential business information submitted by an applicant shall receive the same protection it would receive under state law if reviewed without AI assistance.

Section 9. Transparency, Performance Metrics and Reporting.

(A) The agency shall post a plain-language notice on its permitting website and application forms. The notice shall explain that AI-assisted tools may be used in review and that a qualified reviewer makes every final decision.

(B) Before the pilot begins, the agency shall record baseline figures for each participating program, including median and average processing times, the share of applications returned as incomplete, the number of review cycles per application and the application backlog.

(C) Beginning [date], and every quarter until the pilot ends, the agency shall report to [the appropriate legislative oversight committees and the fiscal research office]. Each report shall include:

  1. A summary of each contract awarded under this Act, including the vendor, term and cost.
  2. A description of the platform’s capabilities and how each participating program used it.
  3. Changes in the baseline figures in subsection (B).
  4. The rate at which qualified reviewers accepted, modified or rejected AI-generated findings, and any identified errors.
  5. An itemized estimate of the pilot’s costs and any staff time saved.
  6. Lessons learned, and recommendations for expanding, modifying or ending the program, including cost estimates, staffing needs and any statutory changes required.

(D) Within [90] days after the pilot ends, the agency shall submit a final report with the information in subsection (C).

Section 10. Funding.

The agency may use funds appropriated for this purpose, existing permit fee revenue [to the extent allowed by law], and available federal or private grants to carry out this Act.

Section 11. Severability.

If any provision of this Act or its application is held invalid, the invalidity does not affect other provisions or applications that can be given effect without the invalid provision or application. To this end, the provisions of this Act are severable.

Section 12. Effective Date

This Act takes effect [July 1, 20XX / upon becoming law].