Critical Questions Districts and IEP Teams Must Address

Artificial intelligence (AI) tools are rapidly finding their way into special education practice. Teachers and administrators, often overwhelmed by documentation demands and growing caseloads, may view AI as a promising way to reduce paperwork and generate present levels of academic achievement and functional performance (PLAAFP) statements, annual goals, accommodations, and recommendations more efficiently.

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.

The use of AI in special education raises significant legal, educational, and ethical questions. AI-generated statements may appear professional and educationally sound while failing to meet IDEA requirements. Goals may not be measurable, may not align with present levels, may omit essential components, or may fail to address the student’s disability-related needs. Similarly, AI-generated recommendations may conflict with federal requirements, state regulations, district policies, or evidence-based instructional practices.

Equally important, AI systems lack the ability to truly know a student. They do not observe the student’s performance in classrooms, understand the student’s strengths and interests, appreciate family priorities, or recognize the nuances of the student’s cultural and linguistic background. AI cannot determine whether a particular service, accommodation, or instructional strategy is appropriate for an individual child.

For these reasons, AI should never be used to make educational decisions. Decisions regarding eligibility, placement, services, supports, accommodations, and annual goals must always be made by qualified professionals and the IEP team based upon evaluation data, present levels of performance, professional judgment, and meaningful team discussion.

The use of AI also raises significant privacy concerns. Teachers may inadvertently enter personally identifiable information into publicly available AI platforms that retain user prompts and uploaded information. Districts must ensure any use of AI complies with the requirements of the Family Educational Rights and Privacy Act (FERPA), IDEA confidentiality provisions, state student privacy laws, and local district policies governing student data.

Perhaps most importantly, the existence of AI-generated content does not transfer responsibility away from educators. The school district and IEP team remain fully responsible for every statement contained in the IEP. If challenged in mediation, a state complaint, or a due process hearing, educators—not the algorithm—must be prepared to explain and defend every goal, service, and decision contained in the student’s program.

Accordingly, whenever AI is used to assist in drafting IEP components, districts should carefully consider the following questions.

Guiding Principles for the Use of AI in IEP Development

1. AI may assist with drafting language but may not replace professional judgment.
2. AI-generated content must be reviewed, revised, and individualized by the IEP team.
3. All PLAAFP statements, goals, services, and accommodations must comply with federal law, state regulations, and district policies.
4. Educational decisions must be made by people, not algorithms.
5. Personally identifiable student information should never be entered into an AI platform unless the district has verified that the platform complies with applicable privacy requirements.
6. The district remains fully accountable for the contents of every IEP, regardless of whether AI assisted in drafting portions of the document.

The Defensibility Test

If the student’s name were removed from the IEP, would an experienced educator be able to determine that this goal was written specifically for this student and no other? If the answer is no, the team should reconsider whether the goal is sufficiently individualized, educationally meaningful, and legally defensible.

Questions Districts and IEP Teams Should Address When Using AI to Assist in Writing IEP Goals

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.

Student Data and Individualization

1. What information was entered into the AI tool?
2. Were current evaluations, observations, and progress-monitoring data considered?
3. Were the student’s strengths, interests, and parent concerns considered?
4. Did the AI-generated language accurately reflect the student’s unique needs?
5. Could this same goal have been generated for many other students?
6. Does the goal address a disability-related educational need?
7. Is the goal necessary for the student to make appropriate educational progress?
8. Would a parent recognize this goal as being written specifically for their child?

Alignment with Present Levels

9. Is there a direct line from the present levels to the proposed goal?
10. What information in the PLAAFP statement supports the need for this goal?
11. Does the goal address an identified educational need rather than simply improve a test score?
12. Does the goal focus on an observable and teachable skill?
13. Does the goal support participation and progress in the general education curriculum?
14. If this goal were removed, what important need would remain unaddressed?

Compliance with IDEA Requirements

15. Is the goal measurable?
16. Does the goal contain observable behavior?
17. Does the goal specify the conditions under which the skill will occur?
18. Does the goal contain objective criteria for mastery?
19. Could an independent observer determine whether the student met the goal?
20. Does the goal avoid vague terms such as “understand,” “improve,” or “demonstrate awareness”?
21. Is the goal ambitious in light of the student’s circumstances, consistent with Endrew F.?
22. Would the district be comfortable defending this goal during mediation or due process?

Specially Designed Instruction

23. What specially designed instruction is necessary for the student to achieve the goal?
24. Does the goal assume interventions that are not documented elsewhere in the IEP?
25. Is there alignment among the student’s need, the goal, and the specially designed instruction?
26. Has the team identified who will provide the instruction and how often?
27. Could another educator implement the instructional plan based solely on the information in the IEP?

Progress Monitoring

28. How will progress toward the goal be measured?
29. What specific data will be collected?
30. Who will collect the data and how often?
31. How often will progress be reported to parents?
32. What decision rules will be used if the student is not making expected progress?
33. Could another educator collect the data and reach similar conclusions?

Professional Judgment and Verification

34. Who reviewed the AI-generated output before it was included in the IEP?
35. What edits were made to the AI-generated language?
36. Can the teacher explain why each component of the goal was retained or changed?
37. Was professional judgment exercised, or was the AI output accepted without sufficient analysis?
38. Would the teacher be comfortable explaining and defending the goal under oath?
39. If the AI system generated a different recommendation tomorrow, would the team’s decision change?

Equity and Bias

40. Could the AI-generated goal inadvertently lower expectations for certain groups of students?
41. Does the goal reflect high expectations and access to grade-level standards whenever appropriate?
42. Were the student’s cultural and linguistic characteristics appropriately considered?

Confidentiality and Privacy

43. Was personally identifiable information entered into the AI system?
44. Does the AI platform retain prompts or uploaded information?
45. Is the use of the AI platform consistent with district policy and applicable privacy laws, including IDEA and FERPA?
46. Have staff received guidance regarding what information may and may not be entered into AI systems?

Documentation and Accountability

47. Does the district have a policy governing AI use in IEP development?
48. Who is ultimately responsible for the contents of the IEP?
49. How will administrators review AI-assisted goals for quality and compliance?

The Ultimate Question

50. If the student’s name were removed from the IEP, would an experienced educator be able to tell that this goal was written specifically for this student and no other?