1. Purpose

This policy establishes guidelines for the responsible use of artificial intelligence (AI) tools by [COMMUNITY NAME] employees, contractors, and volunteers. Its purpose is to ensure that AI is used in ways that protect residents, maintain public trust, support staff, and align with community values.

2. Scope

This policy applies to:

• All employees, contractors, and volunteers of [COMMUNITY NAME]
• All AI tools used for [COMMUNITY NAME] business, including:
  • Commercially licensed AI products (e.g., Microsoft Copilot, ChatGPT Enterprise)
  • Free or consumer AI tools (e.g., ChatGPT free tier, Google Gemini)
  • AI features embedded in existing software (e.g., AI features in Tyler Technologies, Salesforce, or other platforms)

3. Definitions

Artificial Intelligence (AI): Software systems that generate text, images, translations, summaries, predictions, or recommendations. This includes generative AI tools (which create new content) and predictive tools (which analyze data to forecast outcomes or classify information).

Tier 1 — Internal Productivity: AI used by staff for internal work—drafting documents, summarizing reports, research assistance, translation. Output is reviewed by a human before external use.

Tier 2 — Resident-Facing: AI that interacts directly with residents—chatbots, automated responses, public-facing translations. Requires additional governance.

Tier 3 — Decision Support: AI that influences or makes decisions affecting individual rights or access to services. [COMMUNITY NAME] [does not currently use / requires [TITLE] approval for] Tier 3 applications.

4. Permissible Uses

The following uses are approved:

• Drafting documents, reports, correspondence, and internal communications (with human review before use)
• Research assistance and background information gathering
• Summarizing meeting recordings or lengthy documents
• Translation of routine communications
• [Add any community-specific approved uses]

The following uses require approval from [TITLE] before first use:

• Any new AI tool not previously approved
• Any resident-facing AI application (Tier 2)
• Any use involving data described in Section 5 as prohibited
• [Add any community-specific consultation requirements]

The following uses are prohibited:

• Using AI for final decisions affecting resident rights, benefits, or access to services without human review and approval
• Entering information described in Section 5 into consumer or free AI tools
• Representing AI-generated content as original human work without disclosure
• [Add any community-specific prohibitions]

5. Data Handling

The rule: If you wouldn’t share it at a public meeting, don’t enter it into an AI tool.

The following information must never be entered into consumer or free AI tools:

• Personally identifiable information (names, addresses, Social Security numbers, dates of birth)
• Protected health information
• Law enforcement or investigation records
• Employee personnel records
• Attorney-client privileged communications
• Financial account information (bank accounts, credit card numbers)
• Passwords, access credentials, or security configurations
• [Add any community-specific restricted data categories]

For enterprise AI tools (commercially licensed tools configured for government use with data protection agreements): [COMMUNITY NAME] may permit use with [restricted data categories] subject to the vendor’s data handling agreement. [TITLE] must approve such use in writing.

All information entered into AI tools may be subject to public records requests under [state open records law reference].

6. Human Review

All AI-generated output must be reviewed by the responsible employee before use.

For Tier 1 (internal) uses: The employee using the AI tool is responsible for reviewing, verifying, and correcting all output before it is used or shared.

For Tier 2 (resident-facing) uses: [TITLE/POSITION] must review and approve the system configuration, and [frequency] accuracy audits must be conducted.

For any document or decision that affects residents: A supervisor must review AI-assisted output before it is finalized.

Statements of fact in AI-generated content must be verified against primary sources. AI tools may not be cited as a source.

7. Transparency and Disclosure

Internal communications: Employees should note when AI tools were used to draft or substantially assist with work products shared within the organization.

External and public communications: Any document, communication, or decision that represents official [COMMUNITY NAME] position and was substantially assisted by AI must include disclosure. Suggested language:

8. Accountability

Employees who use AI tools are responsible for the output of those tools, just as they are responsible for any other work product.

[COMMUNITY NAME] is responsible for providing employees with the training, time, and resources needed to use AI tools responsibly and review AI output effectively.

Reporting problems: Employees who discover that an AI tool has produced incorrect, biased, or harmful output must report it to [TITLE] immediately. No employee will be penalized for reporting AI errors.

9. Training

Before using any AI tool for [COMMUNITY NAME] business, employees must complete [describe training—e.g., “a 90-minute orientation covering this policy, basic AI capabilities and limitations, data handling requirements, and the review process”].

[TITLE] is responsible for ensuring training is provided and documented.

10. Review and Revision

This policy will be reviewed by [TITLE] every [6 months / annually]. Revisions will be presented to [council / board / governing body] for adoption.

Emergency revisions may be issued by [TITLE] in response to significant changes in technology, state or federal law, or identified governance gaps. Emergency revisions will be presented to [governing body] at the next regular meeting.

11. Effective Date

This policy is effective upon adoption by [governing body] on [DATE].