MODEL NONPROFIT POLICY: USE OF ARTIFICIAL INTELLIGENCE MAIN model for local adaptation, version 2026-10-27 [Organization name] Policy [number]: Use of Artificial Intelligence Approved by [board of directors] on [date], recorded in the minutes. Last reviewed [date]. DRAFTING NOTES (delete before adoption) - This is a model for adaptation. It is not an adopted policy, legal or tax advice, or a requirement from MAIN or any state or federal agency. - Text in brackets marks a decision your organization must make or a name to fill in. Sections 9 and 11 depend on whether you screen people for services and whether you receive grants; remove any section that does not apply. An organization with a few employees and no federal funding needs a much shorter policy than a federally funded service provider. Have counsel review the final draft before the board approves it. - Drafting notes explain options. Delete them before adoption. 1. PURPOSE [Organization name] (the Organization) uses artificial intelligence (AI) to help staff and volunteers serve clients, raise funds, manage grants, communicate, and run operations. AI use also creates risks to the privacy of clients and donors, fair treatment in services and employment, the accuracy of what the Organization tells funders and the public, and security. This policy sets the rules for using AI in the Organization's work. Procedures issued under this policy set the details. 2. SCOPE This policy applies to board members, employees, volunteers, [LOCAL DECISION: interns and contractors], and vendors acting for the Organization when they use AI for the Organization's work, on its devices or accounts, or with its information. It covers stand-alone AI tools, AI features inside software the Organization uses, and AI agents. Volunteer and contractor agreements will refer to this policy. 3. DEFINITIONS a. Artificial intelligence. [LOCAL DECISION: Use the definition in Mississippi law, shown here, or adopt your own.] "Artificial intelligence" means a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations or decisions influencing real or virtual environments. Artificial intelligence systems use machine- and human-based inputs to perceive real and virtual environments; abstract such perceptions into models through analysis in an automated manner; and use model inference to formulate options for information or action. (House Bill 1723, 2026 Regular Session, codified in Chapter 3, Title 1, Mississippi Code of 1972.) b. AI tool means any software, service, or product feature that uses artificial intelligence to generate content, analyze information, make predictions or recommendations, or take actions. c. AI agent means an AI tool that can take actions in other systems on its own, such as sending messages, changing records, submitting forms, or making payments. d. Approved AI tool means an AI tool the Organization has approved under Section 7 for a stated purpose, group of users, and category of information. e. Restricted information means information about clients and program participants, including names, contact details, case notes, health information, and immigration or financial details; donor personal and giving information; personnel and volunteer records; information about children; Social Security, account, and card numbers; and [LOCAL DECISION: the other categories the Organization designates]. f. High-impact use means any AI use that (1) informs a decision about a person's eligibility for, or continuation of, services or assistance; (2) informs an employment or volunteer selection decision; (3) communicates with clients, donors, or the public without a person reviewing each message first; (4) prepares grant applications, funder reports, or financial statements submitted without review; (5) uses restricted information; or (6) allows an AI agent to act for the Organization. 4. PRINCIPLES The Organization's use of AI will follow these principles: - a person remains accountable for every decision affecting a client, donor, employee, volunteer, or applicant; - AI is used where it advances the Organization's mission or frees capacity for it; - information about clients and donors was given in trust and is entered only into tools the Organization has approved; - review scales with the consequence of the use; - the Organization is willing to tell its board, donors, and clients how it uses AI; and - facts, figures, citations, and names in AI-generated material are verified before they leave the Organization. 5. RESPONSIBILITIES a. The board of directors will approve this policy, review it at least [LOCAL DECISION: once a year], and receive reports on AI use in high-impact areas. b. [The executive director] will carry out this policy, approve AI tools, and authorize exceptions within the limits of law, funder terms, and this policy. c. [The AI point of contact] will keep the list of approved tools, answer questions, and log incidents. d. Program leads will decide whether AI is appropriate in their service area and record the reason. e. Each staff member and volunteer is responsible for any work they produce, send, or approve with AI assistance and must report problems. DRAFTING NOTE: In a small organization the AI point of contact is a role, not a position. The executive director or an operations lead can hold it, with a board member as reviewer. 6. PERMITTED, APPROVED, AND PROHIBITED USES a. Permitted. Staff and volunteers may use approved AI tools, with human review, for drafting internal documents and routine correspondence, summarizing public materials, brainstorming, first drafts of grant narratives, training materials, and translation of published content checked by a qualified person. b. Permitted with approval. Any high-impact use, any use of restricted information, and any AI tool that is not yet approved require approval under Section 7 before use begins. c. Prohibited. No one may use AI for the Organization's work to: (1) enter restricted information into a tool that is not approved for it and covered by an appropriate written agreement; (2) make a final decision on eligibility, benefits, denial or termination of services, hiring, discipline, or dismissal; (3) create images, audio, or video of real people, including clients, donors, staff, or board members, without their written consent; (4) present AI-generated testimonials, client stories, quotations, or impact figures as authentic; (5) present AI-generated text as the personal writing of a named person who did not write or review it; (6) submit AI-generated content to a funder whose terms prohibit it; or (7) [LOCAL DECISION: add any other prohibited uses the Organization chooses]. DRAFTING NOTE: The board may grant an exception only to an internal rule it has authority to change. It cannot waive applicable law, binding funder terms, or the human decisions this policy requires. 7. INVENTORY, REVIEW, AND APPROVAL a. The Organization will keep a written inventory of AI tools in use, including tools staff adopted without approval, showing who approved each, what information may be entered, and the date of last review. b. Before an AI tool is approved for a high-impact use or restricted information, [the executive director or AI point of contact] will obtain written answers to at least these questions: (1) Will the vendor use our data to train its models or anyone else's, and can we refuse by contract rather than a settings toggle? (2) Where is data stored, for how long, and how is it deleted when we leave? (3) Will the vendor sign a business associate agreement or data processing agreement if we need one? (4) How quickly will the vendor tell us about a breach, and does that let us meet our own notice duties? (5) Which subcontractors handle our data? (6) What accessibility does the product support, and is there a current accessibility conformance report? (7) For an AI agent, what actions can it take, what are its limits, how are its actions logged, and who can stop it? c. This review also applies when software the Organization uses adds AI features, and again after any material change. d. Approval may be limited, paused, or withdrawn when a tool does not perform as expected, causes harm, or no longer meets these requirements. 8. CLIENT AND DONOR DATA, PRIVACY, AND SECURITY a. Restricted information may be entered only into approved tools covered by a written agreement on confidentiality and training use, and only the minimum necessary for the task. b. Protected health information may be entered only into a tool whose vendor has signed a business associate agreement, if the Organization is a HIPAA covered entity or business associate. c. Student records shared with the Organization by a school or college will be used only as the sharing agreement allows. d. Files containing Social Security, driver's license, or financial account numbers will be encrypted. e. Because AI can imitate voices, video, and writing, staff must verify unusual requests for payments, banking changes, credentials, or donor or client information through a separate, known contact method before acting on them. 9. PROGRAM SERVICES AND ELIGIBILITY a. [LOCAL DECISION: For organizations that screen people for services, housing, or assistance.] AI used in intake, screening, eligibility, prioritization, or termination of services is a high-impact use. b. A qualified person will make and document each decision to deny, reduce, or end services, keep the basis for the decision, and offer an appeal that reaches a person. c. The Organization will monitor outcomes and complaints for errors and unequal effects on groups of people, and will document how problems are corrected. 10. FUNDRAISING AND DONOR COMMUNICATIONS a. Every statistic, story, and claim in an AI-drafted appeal will be checked against the Organization's records before it is sent. b. The Organization will [LOCAL DECISION: disclose AI-generated or AI-modified images in fundraising materials]. c. [The development lead] will track whether AI-assisted targeting or online giving brings the Organization within charitable solicitation registration rules in Mississippi or other states. 11. GRANTS AND FUNDER REQUIREMENTS a. Before AI is used to prepare an application or report, the grant lead will read the funding notice and award terms for any requirement about AI. b. Grant applications, reports, and budgets prepared with AI assistance will be reviewed and approved by [the responsible staff member] before submission. c. For federal awards, prompts, outputs, and logs that document work under the award will be kept as award records for the required retention period, and AI tools will be included in the Organization's internal controls and safeguards for personally identifiable information. 12. EMPLOYMENT AND VOLUNTEERS a. AI used in recruiting, screening, interviewing, evaluating, or scheduling staff or volunteers is a high-impact use and requires approval by [the executive director] before use. b. A person will make each hiring, discipline, and dismissal decision. c. The Organization will [LOCAL DECISION: tell applicants when AI is used in decisions about them], offer reasonable accommodations, and check for unequal effects. 13. ACCESSIBILITY AND TRANSPARENCY a. AI-generated images will have accurate alternative text, AI-generated video will have accurate captions, and AI tools on the Organization's website or apps will be checked for accessibility before launch. b. AI chat assistants on the Organization's website will be labeled as automated and will offer a way to reach a person. c. The Organization's privacy notice will [LOCAL DECISION: state whether client or donor information is processed using AI tools]. 14. VENDORS, SOFTWARE, AND AI AGENTS a. AI tools bought with federal funds will follow the Organization's procurement standards for federal awards. b. If a board member, officer, or their family has an interest in an AI vendor under consideration, the Organization's conflict of interest process applies. c. Scripts, applications, integrations, websites, low-code workflows, and agents created with AI must be reviewed by [the AI point of contact or a qualified reviewer] before they use restricted information, receive privileged access, are shared, or are relied on for operations. d. An AI agent may act for the Organization only within the actions and limits approved under Section 7. Each agent will have a named person who is responsible for it and can stop it, and an agent may not submit, send, publish, change a record, or make a payment without the approval that would apply to a person. Each proposed consequential action must receive qualified human review and approval before execution; approval of the agent, access, or workflow does not approve its later consequential actions. 15. RECORDS AND INCIDENTS a. The Organization's document retention policy will state how long AI prompts, outputs, and logs are kept. b. Staff and volunteers must report AI-related incidents promptly to [the AI point of contact]. Incidents include restricted information entered into an unapproved tool, inaccurate AI-generated content that reached a donor, client, or funder, a vendor breach, an accessibility failure, and a questionable automated outcome. c. The executive director will be notified promptly. The Organization will contain the issue, decide whether breach notice is required under Mississippi law or notice is owed to a funder or regulator, correct any public-facing error, and log the incident even when no notice is required. 16. TRAINING, COMPLIANCE, AND REVIEW a. Before receiving access to approved tools, staff and volunteers will complete training on what may never be entered into an AI tool, who approves a new tool, what must be verified before publication, and how to report a problem. The Organization will record who completed training and when. b. Violations of this policy may result in [LOCAL DECISION: action under the Organization's personnel and volunteer policies]. Staff and volunteers may raise concerns about AI use in good faith without retaliation. c. The board will review this policy at least [LOCAL DECISION: once a year] and when funding sources, employee count, services, or the tool inventory change materially. LEGAL REFERENCES Legal references. [LOCAL DECISION: Confirm with counsel and keep only those that apply to your organization's funding, size, and services.] - Title VI of the Civil Rights Act of 1964, 42 U.S.C. § 2000d et seq. (federal financial assistance) - Section 504 of the Rehabilitation Act, 29 U.S.C. § 794; Age Discrimination Act of 1975, 42 U.S.C. § 6101 et seq. (federal financial assistance) - Americans with Disabilities Act, Title I, 42 U.S.C. § 12111 et seq., and Title III, 42 U.S.C. § 12181 et seq. - Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq.; Age Discrimination in Employment Act, 29 U.S.C. § 621 et seq. - Fair Housing Act, 42 U.S.C. § 3601 et seq. (housing-related activities) - HIPAA, 45 C.F.R. Parts 160 and 164 (covered entities and business associates); FERPA, 34 C.F.R. Part 99 (student records shared by schools) - Uniform Guidance, 2 C.F.R. Part 200, including §§ 200.303, 200.332, and 200.334 (federal awards) - Section 75-24-29, Mississippi Code of 1972 (notice of breach of security) - Mississippi Charitable Solicitations Act, Section 79-11-501 et seq., and 1 Miss. Admin. Code Pt. 15 - Mississippi Nonprofit Corporation Act, including Sections 79-11-267 and 79-11-269 - House Bill 1723, 2026 Regular Session (definition of artificial intelligence) Cross-references: [privacy notice; data classification; conflict of interest policy; whistleblower policy; document retention and destruction policy; personnel and volunteer policies; procurement; incident response; website accessibility] Source: Mississippi Artificial Intelligence Network (MAIN), Nonprofit AI Policy Template, https://mainms.org/ai-policy-and-guidance-template-nonprofit/. General information, not legal or tax advice. See https://mainms.org/legal-information/.