Students Come First

AI should support each student's individual IEP goals, strengths, and needs. In the same ways AI can be used as a tool to support student academic and functional growth; AI can also be used as a tool to harm students. Teachers are responsible for providing education to all students on how AI can be used to create false information such as through advertising or videos that may cause financial or emotional harm. AI education is especially important for students who are more susceptible to being influenced by deceptive information or individuals that may being trying to take advantage of others.

Accessibility by Design

AI should improve access to learning while complementing existing accessibility and assistive technology support.

Human Oversight and Professional Judgment Matters

Educators, related service providers, and families remain responsible for educational decisions and individualized supports. Professional judgement and decision making are especially important and required when developing a student’s Individualized Education Program (IEP). The Human-AI-Human framework ensures that human expertise and oversight remain central to the process, with AI serving as a supportive tool. 

The Human-AI-Human (H-AI-H) Workflow

1. Human Inquiry: The process begins with human-initiated prompts based on anonymized comprehensive student data, contains no Personally Identifiable Information (PII), and maintains professional judgement.

2. AI Assistance: AI provides initial drafts and suggestions as a support tool such as for leveling or summarizing text, pre-teaching vocabulary, or creating study questions that are then reviewed and edited by the person responsible for developing content.

3. Human Review and Finalization: Educators critically review, revise, and ensure all content meets individualized needs and legal standards.

Protect Student Privacy

AI use should align with district policies and applicable privacy requirements while safeguarding confidential student information. In addition, families should be cautious in using AI when communicating important information about their child such as through an IDEA State Complaint. Use only secure, district approved devices and networks when accessing AI. Check compliance of AI platforms with Family Educational Rights and Privacy Act (FERPA) and Health Insurance Portability and Accountability Act (HIPAA) as well as district policies.

Equity and Inclusion

Be aware that AI algorithms and models can inadvertently reflect biases and disparities. Educators should proactively mitigate bias or disparities by monitoring them. Use inclusive, representative, and objective input when prompting AI tools. Check for subjective statements about students as well as cultural, linguistic, disability, gender, or racial bias.

Professional Ethics

• Align the use of AI with professional codes (e.g., American Speech-Language-Hearing Association (ASHA), American Occupational Therapy Association (AOTA), American Physical Therapy Association (APTA), National Association of School Psychologists (NASP), National Association of Social Workers (NASW), or National Association of School Nurses (NASN).

• Do not delegate clinical judgment to AI, use it only as a support tool.

• Cross-reference AI outputs with current research and evidence-based best practices to ensure alignment with effective instructional strategies (e.g., checking resources from professional organizations, academic journals, higher education institutes or federally funded technical assistance centers).

Legal Precedent and Cautions

While AI may be a helpful tool in limited situations, a district must be cautious about utilizing it for a number of reasons, including the following:

• In at least one case, a court determined that a district denied a student Free and Appropriate Public Education (FAPE), because the district relied on technology to bypass the student’s learning gaps regarding reading; instead of attempting to use instructional strategies to teach the student foundational reading skills. (Source: W.A. v. Clarksville-Montgomery County Sch. Sys., 125 LRP 3627, 127 F.4th 656 (6th Cir. 2025)).
• Personally identifiable student information may never be entered into generative AI or search engines. This includes any information that may indirectly identify the student, such as their disability, grade level, race, etc. “AI hallucination is a phenomenon where, in a large language model (LLM) often a generative AI chatbot or computer vision tool, perceives patterns or objects that are nonexistent or imperceptible to human observers, creating outputs that are nonsensical or altogether inaccurate.” Simply stated, AI Hallucinations are false information presented as facts. For example, a court case with a finding favorable to an attorney’s case that does not actually exist. “Generally, if a user makes a request of a generative AI tool, they desire an output that appropriately addresses the prompt (that is, a correct answer to a question).”  For more information see What are AI Hallucinations. 

Transparency and Disclosure

Any use of AI should be disclosed when applicable to team members and families. Families should be informed of their role in education and their child’s Individualized Education Program (IEP). Ensure families understand that human professionals maintain full responsibility for IEP content. In addition, conduct annual compliance audits to verify proper implementation of ethical, legal, and functional AI practices.

Individualized Education Program (IEP)

• Generating initial information about student performance from educator-provided student data (ensuring any personal identifiable information (PII) removed before entering data into AI platform).

• Drafting SMART goal suggestions aligned with assessments and standards.

• Rephrasing academic and functional performance descriptions for family-friendly language.

Note that AI may assist in drafting language, but it cannot exercise professional judgment, determine educational need, make eligibility or placement decisions, select services, or replace the individualized decision-making responsibilities of the IEP team. The district remains fully responsible for every statement, goal, service, and decision contained in the IEP. 

Used thoughtfully, AI may serve as a brainstorming tool or provide a starting point for drafting language. However, AI cannot replace professional judgment, individualized decision-making, or the collaborative responsibilities of an IEP team.

Instruction

• Lesson Planning: Generating differentiated lesson plans aligned with standards.

• Differentiated Activities: Creating varied activities around common themes to meet diverse learning needs.

• IEP Implementation Support: Generating instructional materials that align with student needs and goals.

• Universal Design for Learning (UDL) Alignment: Supporting differentiated instruction by generating alternative lesson formats, adapting content to different learning preferences, and providing multimodal access to instruction.

• Task Analysis and Breakdown: AI can help break down complex tasks into discrete steps for instruction and creating visual supports for sequencing tasks.

Data and Communication

• Data Analysis: Identifying patterns in student performance data and creating visuals of the data.

• Multilingual Support: Providing translations for students and families with diverse language backgrounds (It is important to note that AI translations are not always 100% accurate and are not a replacement for existing translation services and practices). AI translations tools can be helpful for general information and for routine conversations.  But AI translations may contain significant errors, omissions and details or misinterpreted educational, legal or culturally specific language.   Suggested use may include assisting in drafting or translating informational materials. However, AI translation should not replace qualified human translators and interpreters. When communications about special education services, eligibility, evaluation, IEPs, and other legal requirements that involve legal rights and educational decision making, schools must ensure that families have information in a language they understand. 

• Administrative Support: Assisting with clerical tasks and administrative documentation.

• Scheduling: Assisting in managing session scheduling efficiently.

• Suggesting Evidence-based Strategies and Interventions: All practices must be cross-checked for accuracy of evidence for the specific student demographics and unique student’s disability-related needs.

(Adapted from North Carolina DPI Technical Assistance: Acceptable Use of AI for Special Education Teachers and Related Service Providers, 2025)