City AI Governance ProjectA Civic AI Lab project

State guide

Florida AI Governance Guide for Local Governments

A practical starting point for Florida cities, counties, school districts, special districts, and local agencies evaluating, purchasing, or using AI-enabled systems.

Version 1.0Source-verified: July 16, 2026Policy and governance resource — not legal adviceComprehensive 2026 AI proposals: not enacted

Disclaimer

This resource is for policy and governance education only. It is not legal advice. Local governments should consult their attorney or legal counsel before making procurement, compliance, or deployment decisions.

← All state guides
On this page

Florida does not currently have one comprehensive municipal AI governance code.

As of July 16, 2026, Florida has no enacted statute creating one comprehensive governance framework for every municipal use of artificial intelligence. Existing duties apply according to the agency, function, contract, record, communication, data, and decision involved.

A local government should apply those duties to a specific system through its normal legal, records, procurement, IT/security, and departmental review processes.

State AI policy context

Florida’s state-level AI framework

Florida section 282.802 establishes the Government Technology Modernization Council. The council studies new technologies and recommends approaches to procurement and regulation; its AI-related duties include considering a state AI code of ethics for state government, assessing automated decision systems and legal rights, evaluating AI safety and security standards, and examining governmental and private-sector AI use.[13]

Chapter 2026-66 amended section 282.802 only to conform a statutory cross-reference concerning foreign countries of concern. It did not create a comprehensive municipal AI governance program or new general local-government AI compliance regime.[14]

Local-government meaning

The council is relevant for monitoring future Florida policy. Its existence does not itself require a city or county to adopt a particular AI inventory, risk framework, notice, impact assessment, or procurement process.

At a glance

Five things Florida local officials should know

Public records remain central

Florida policy states that state, county, and municipal records are open for inspection and that automation, electronic systems, and proprietary software must not erode lawful access to public records.[1]

AI records classified by role in official business

Florida defines public records broadly as material made or received in accordance with law or in connection with an agency’s official business, regardless of physical form or means of transmission. An AI prompt, output, log, recommendation, or review note falls within that definition when it meets that official-business test; disclosure and retention then depend on applicable exemptions and records schedules.[2]

Some vendor relationships carry public-records duties

When a vendor qualifies as a contractor acting on behalf of a public agency under section 119.0701, covered contracts must address maintenance, production, protection, transfer, and compatible electronic delivery of relevant public records.[3]

Election-related AI disclosure is narrow but real

Florida requires a specified disclaimer for certain political advertisements or political communications containing generative-AI material that depicts a real person performing an action that did not occur and meets the statute’s intent requirements.[4]

Broader 2026 AI proposals did not become law

CS/SB 482 died in House messages, SB 2-D died in a House subcommittee during the 2026 special session, and SB 146 died in a Senate committee. Their proposed requirements are not current law.[5][6][7]

What applies now — 1

Public records and AI-assisted government work

Florida’s public-records policy applies to municipal as well as state and county records. The statute states that automation must not erode public access, agencies must consider access when acquiring electronic recordkeeping systems, and proprietary software must not diminish lawful inspection and copying rights.[1]

Florida’s records-management guidance defines a public record by its relationship to official agency business, regardless of physical format or means of transmission. It also explains that retention is determined by the record’s nature, content, and purpose.[2]

What this may mean for AI

An AI prompt, output, transcript, recommendation, audit log, or human-review record falls within Florida’s public-record definition when it is made or received in accordance with law or in connection with official agency business. That is an application of Florida’s general records definition, not an AI-specific statutory rule. Do not state that every prompt or output is automatically a public record.

Questions local officials should ask

  • Is the AI-assisted material connected to official agency business?
  • Did it inform a communication, recommendation, permit, employment action, enforcement action, benefit, or other decision?
  • Is the record stored only inside a vendor platform?
  • Can the agency search, export, and produce the relevant information?
  • Which retention schedule applies?
  • Does an exemption or confidentiality provision apply?
  • Is documentation sufficient to understand the system's role and the final human decision?
  • Can records be preserved during litigation, an audit, an investigation, or a pending records request?

Recommended governance action

Create interim internal guidance that explains:

  1. 01when staff should preserve AI-assisted work;
  2. 02what metadata or logs should accompany consequential uses;
  3. 03which office decides the applicable retention schedule;
  4. 04how records are exported from vendor systems;
  5. 05how confidential or exempt information is protected.

What applies now — 2

Records retention and disposal

The Florida Department of State publishes general records schedules for state and local agencies and separate schedules for criminal justice, election, public-school, fire, utility, and other functions. The Department explains that not every record is covered by a general schedule and directs agencies to work with their Records Management Liaison Officer when a record series is unclear.[8][2]

Practical implication

Do not create an arbitrary “AI records retention period” for every use.

Instead, classify AI-assisted records according to:

  • the government function they document;
  • the underlying decision or transaction;
  • the department that maintains them;
  • the applicable general or function-specific schedule;
  • litigation, audit, investigation, and public-records constraints.

Minimum inventory fields

For every AI-enabled system, record:

  • records created;
  • records received;
  • prompts and outputs retained;
  • logs available;
  • location of storage;
  • export format;
  • applicable retention schedule;
  • records owner;
  • vendor obligations;
  • termination and migration process.

What applies now — 3

Vendor and contractor records

Section 119.0701 applies to a defined category of contractors: entities entering into contracts for services with a public agency and acting on behalf of the agency as described by law. It should not be described as automatically applying to every software vendor or every technology purchase.[3]

For covered contracts, the statute addresses:

  • maintaining required public records;
  • providing records to the agency or allowing inspection and copying;
  • protecting confidential and exempt records;
  • transferring or maintaining records after contract completion;
  • providing electronic records in a format compatible with the agency's systems.

Why this matters for AI procurement

AI-enabled products can place important records inside vendor-controlled systems. A local government may need access to:

  • prompts and outputs;
  • system and model versions;
  • decision-support recommendations;
  • user activity;
  • logs;
  • human actions;
  • notices;
  • complaints;
  • corrections;
  • incident records.

Important qualification

Whether section 119.0701 applies to a specific vendor depends on the contract and the vendor’s role. The city or county attorney should make that determination.

Even when the section does not directly apply, similar access, retention, export, and termination provisions may be prudent as contract safeguards.

What applies now — 4

Elections and synthetic political media

Florida section 106.145 applies to specified political advertisements, electioneering communications, and other political advertising when all statutory conditions are met. The law covers content created in whole or in part with generative AI that appears to depict a real person performing an action that did not occur and was created with the specified intent to injure a candidate or deceive regarding a ballot issue.[4]

The statute prescribes the disclaimer:

“Created in whole or in part with the use of generative artificial intelligence (AI).”

It also specifies format requirements for print, video, online, audio, and graphic communications.[4]

What this does not mean

This is not a general labeling requirement for every local-government use of AI.

Local election officials, public-information staff, candidates, campaigns, and counsel should distinguish:

  • official government communications;
  • political advertising;
  • electioneering communications;
  • ballot-issue communications;
  • synthetic media that does not satisfy every statutory condition.

Questions to review

  • Is the communication political or election-related?
  • Does it depict a real person performing an action that did not occur?
  • Does the statutory intent requirement apply?
  • Who is paying for, sponsoring, or approving the communication?
  • What format-specific disclaimer is required?
  • Are other campaign-finance or election-law disclaimers also required?

What applies now — 5

School districts and educational AI

Florida section 1002.321 permits school districts to receive grant funding for certain AI platforms supporting students in grades 6 through 12. For grant recipients, the selected platform must satisfy specified criteria, including use of a closed system, teacher professional learning, standards-aligned functions, reporting, and parental access to AI interactions.[9]

This provision is not a comprehensive school-district AI governance code. It is a use-specific statute connected to a grant program.

School-district governance questions

  • Does the tool process student records or sensitive information?
  • Is the system closed, and what does the vendor mean by “closed”?
  • Can parents access relevant interactions where required?
  • How are inaccurate or harmful outputs corrected?
  • What human review is provided by teachers or staff?
  • What records are retained?
  • Can the vendor use student interactions to train or improve unrelated models?
  • Are accessibility and language-access requirements met?
  • Are student, parent, teacher, and employee notices clear?
  • Which state and federal student-data requirements also apply?

Governance recommendation

Recommended enhanced review for high-risk uses

This section is a governance recommendation.

It is not a statement that Florida law creates one universal review procedure for every listed system.

Apply enhanced legal, records, privacy, civil-rights, security, procurement, and operational review to systems involving:

  • policing and surveillance;
  • facial recognition and biometric identification;
  • automated license-plate readers and video analytics;
  • emergency dispatch and public safety;
  • employment screening, evaluation, discipline, or scheduling;
  • benefits and access to essential services;
  • permits, inspections, code enforcement, and fraud detection;
  • housing;
  • education and student data;
  • systems whose outputs strongly influence liberty, safety, rights, or livelihood.

Enhanced-review questions

  • What decision does the system influence?
  • What authority remains with a human official?
  • Can the human reviewer meaningfully change the result?
  • What data is used?
  • Who may be affected?
  • How are errors detected?
  • How can a person seek correction or review?
  • What audit evidence exists?
  • What records are created?
  • Can the city suspend the system after an incident or material vendor change?

Failed proposals

What did not become law

CS/SB 482 — Artificial Intelligence Bill of Rights — Did not pass

The Florida Senate passed CS/SB 482 on March 4, 2026. The official history states that it died in House messages on March 13, 2026.[5]

Do not present its proposed consumer, procurement, chatbot, education, or other provisions as current law.

SB 2-D — Artificial Intelligence Bill of Rights, 2026 Special Session — Did not pass

The Senate passed SB 2-D on April 28, 2026. The official history states that it died in the House Information Technology Budget & Policy Subcommittee on April 29, 2026.[6]

Do not present its proposed local-government contracting requirements as enacted obligations.

SB 146 — Use of Artificial Intelligence by State Agencies — Did not pass

SB 146 would have required the Florida Digital Service to study state-agency AI procurement and use. It died in the Senate Governmental Oversight and Accountability Committee on March 13, 2026.[7]

It concerned state agencies and did not become law.

Why this section matters

Articles, summaries, presentations, and vendor materials may continue to discuss these proposals. Local officials should confirm the official bill status before treating any proposed provision as a current requirement.

Florida local examples

Illustrative local practices

Local example — not a statewide requirement

Miami-Dade County: employee AI policy and governance structure

Miami-Dade County publicly describes policies requiring employees to use approved generative-AI tools, protect sensitive information, conduct human review and validation, disclose AI use in public-facing work as appropriate, complete training, and report concerning outputs. The County also publishes AI governance materials and examples of priority AI projects.[10][11]

Practical lessons

  • maintain an approved-tools process;
  • prohibit sensitive data in unapproved public tools;
  • require human review;
  • assign a reporting channel for incidents;
  • connect use cases to an IT-led governance process;
  • publish enough information for residents to understand how the county is using AI.

Do not describe Miami-Dade’s policy as a statewide requirement.

Local example — not a statewide requirement

Orlando Connect: public disclosure for AI-assisted video analytics

The Orlando Connect project page states that Axon Fusus uses AI to search video provided to the system, that the disclosed AI uses exclude facial recognition, and that access to connected cameras requires authorization. The same page publishes security and hosting claims about the platform.[12]

Practical lessons

  • explain what the system does;
  • state what it does not do;
  • describe authorization and access;
  • provide security and data-handling information;
  • avoid treating a vendor label such as “AI” as sufficient public explanation.

This example is not an endorsement or an independent verification of the platform’s security, performance, or legal compliance.

Recommended first 90 days

A practical starting sequence

Days 1–30: Map and assign responsibility

  1. 01Inventory AI-enabled tools across departments.
  2. 02Include tools embedded inside existing software.
  3. 03Record where prompts, outputs, logs, and recommendations are stored.
  4. 04Identify the records owner and internal system owner.
  5. 05Assign a named official or cross-department working group.
  6. 06Flag systems involving public safety, employment, education, benefits, housing, permits, biometrics, or surveillance.

Days 31–60: Review records and contracts

  1. 01Ask the Records Management Liaison Officer to map relevant schedules.
  2. 02Review whether vendor-held records can be searched and exported.
  3. 03Determine which vendors may qualify as contractors acting on behalf of the agency.
  4. 04Review data reuse, model changes, subcontractors, audit access, incident reporting, and termination.
  5. 05Identify renewals or amendments that may introduce new AI features.
  6. 06Create interim staff guidance for approved tools and sensitive information.

Days 61–90: Establish oversight

  1. 01Assign preliminary risk tiers.
  2. 02Create a review pathway for high-risk systems.
  3. 03Establish human-review and correction procedures.
  4. 04Create incident-reporting and suspension procedures.
  5. 05Prepare resident-facing disclosure for relevant systems.
  6. 06Train staff.
  7. 07Schedule an inventory refresh and contract review.

For counsel and records officers

Questions for the city or county attorney and records officer

  1. 01When does an AI prompt, output, transcript, recommendation, or log document official agency business?
  2. 02Which retention schedule applies to each category?
  3. 03What documentation should accompany a consequential AI-assisted decision?
  4. 04Which exemptions or confidentiality provisions may apply?
  5. 05Which vendors qualify as contractors acting on behalf of the agency under section 119.0701?
  6. 06Can the agency retrieve relevant records directly, or must it depend on the vendor?
  7. 07Are records available in a compatible and usable electronic format?
  8. 08What must be preserved during litigation, an audit, an investigation, or an active records request?
  9. 09What public notice or disclosure is legally required for this use?
  10. 10What administrative review, appeal, due-process, employment, civil-rights, education, or sector-specific duties apply?

Florida-specific procurement questions

Questions to work through before buying or renewing

  1. 01AI features — Does the product include AI, machine learning, generative AI, automated ranking, prediction, detection, transcription, or decision-support functions? Identify each base, optional, and planned feature.
  2. 02Records created — What prompts, outputs, logs, recommendations, notices, corrections, and human actions does the system create or retain?
  3. 03Official-business records — Which records may document official agency business, and who controls them?
  4. 04Search and export — Can the agency search and export relevant records without vendor assistance? In what formats?
  5. 05Compatible delivery — Can electronic records be supplied in a format compatible with agency systems when required?
  6. 06Exempt and confidential information — How does the vendor segregate, protect, and return confidential or exempt records?
  7. 07Retention — Can the system support applicable retention schedules, legal holds, audits, investigations, and pending public-records requests?
  8. 08Contractor status — Is the vendor acting on behalf of the agency in a way that may trigger section 119.0701? Counsel should make this determination.
  9. 09Data use — Will city, employee, student, resident, prompt, output, or derived data be used to train, fine-tune, evaluate, or improve vendor or third-party models?
  10. 10Subcontractors and model providers — Which cloud, model, analytics, and subcontractor entities receive data or provide system functions?
  11. 11Material changes — Can the vendor change the model, provider, data practices, output behavior, or AI features after contract signing? What notice and approval rights apply?
  12. 12Testing and auditability — What performance documentation, testing access, version information, logs, and audit rights are available?
  13. 13High-risk uses — Does the system affect employment, housing, benefits, education, permits, public safety, enforcement, surveillance, or another consequential function?
  14. 14Human authority — Who retains final decision authority, and can that person meaningfully reject or correct the system's recommendation?
  15. 15Incidents — What must the vendor report, within what period, and what records must be preserved?
  16. 16Suspension — Can the agency disable an AI feature without terminating the full software platform?
  17. 17Termination and portability — What data and records are returned, retained, deleted, or certified as destroyed at termination?
  18. 18Public explanation — Can the agency accurately explain the system's purpose, limits, data, human oversight, and complaint process to residents?

Proportionate review

Apply the most detailed review to systems with consequential or high-risk uses. Lower-risk internal tools may receive a proportionate review, but data protection, records access, system changes, and vendor accountability should still be addressed.

Help improve this guide

Corrections and source updates

This guide has been checked against the official sources listed below. Florida records professionals, procurement officials, IT leaders, school-district staff, municipal attorneys, and other public-sector practitioners are invited to submit corrections, newer official sources, or implementation examples.

Submit a correction via the project contact form →

Disclaimer

Scope of this guide

This guide provides general policy and governance information. It is not legal advice, a comprehensive statement of Florida law, or a substitute for review by a city, county, school-district, or agency attorney. Laws, regulations, guidance, records schedules, and bill statuses may change. Officials should verify current requirements and the facts of a particular use before acting.

Official sources

Where to verify each claim

Download

Florida Local AI Governance Checklist

One-page checklist (PDF) adaptable for your city, county, or agency.

Download the Florida Local AI Governance Checklist (PDF)