Navigating AI in IEP Development: A Framework for Ethical Practice
Authors: Olivia Fudge Coleman, Ph.D. and Danielle A. Waterfield; info@ciddl.org
[embedyt] https://www.youtube.com/watch?v=yaSC-J0iAq0[/embedyt]
Late on a Thursday night, Mr. Schrute faces a familiar challenge: translating assessment data and observations into clear, measurable IEP goals for eight students. Exhausted, he experiments with a generative AI (GenAI) tool, pasting in a student summary. Within seconds, he has a draft goal. Relieved by the efficiency, he drops it into the IEP without careful review. However, he fails to notice that it lacks measurability, omits necessary supports, and includes the student's full name in auto-generated language.
This scenario is playing out in schools across the country as special educators increasingly turn to GenAI tools to manage overwhelming workloads. While these tools promise to streamline IEP development and improve goal quality, they also introduce significant ethical concerns around individualization, bias, privacy, and accuracy.
The Promise and the Peril
Research suggests GenAI can be a powerful support when used strategically. Studies by Rakap (2024) and Waterfield et al. (2025) found that educators using ChatGPT with proper guidance produced IEP goals of the same or higher quality while spending less time drafting. Some platforms even claim to reduce paperwork time by up to 90% (eSchool News Staff, 2025).
However, these efficiency gains come with substantial risks:
- Loss of Individualization: AI-generated content often reflects templated language rather than student-specific strengths and needs (Evans & Sinha, 2024; Imada, 2024; Nixon et al., 2024)
- Algorithmic Bias: Large language models trained on non-representative datasets may perpetuate systemic inequities already present in special education (Dodge et al., 2021; Glazko et al., 2024; Kolisko & Anderson, 2023; Ocumpaugh et al., 2024; U.S. Department of Education, 2023).
- Privacy Violations: Many AI platforms store data on external servers, potentially exposing protected student information and violating FERPA (Perez, 2024).
- Hallucinations: AI can generate confident but inaccurate content that appears authoritative, undermining both instructional planning and legal compliance (Jin et al., 2025).
Framework for Ethical AI Use
Rather than rejecting AI tools or adopting them uncritically, we propose a comprehensive framework grounded in legal mandates, professional judgment, and student-centered planning. The framework centers on six core principles:
Principle 1: Professional Responsibility and Judgment
Core Idea: AI output must always be treated as a first draft requiring rigorous human review.
Action Steps:
- Mark all AI-generated content and maintain an edit log documenting every human revision.
- Compare outputs against IDEA requirements for measurable, evidence-based goals.
- Check for compliance issues before sharing with teams or families.
- Never put identifiable information (e.g., full student name, birthdate, school name) into an AI platform.
Principle 2: AI Fluency and Prompt Crafting
Core Idea: Effective prompts generate more accurate, equitable, and measurable content.
Action Steps:
- Develop and regularly update prompt template libraries with placeholders for assessment data.
- Implement quick bias checks using rubrics focused on measurability and cultural responsiveness.
- Engage in regular, brief training sessions (even 15 minutes monthly) on prompt literacy.
- Maintain a "prompt portfolio" documenting versions, outputs, edits, and rationales.
Principle 3: Individualization
Core Idea: Generic goals. Every plan must reflect the individual student.
Action Steps:
- Cross-check AI drafts against recent evaluations, family input, and student preferences.
- Apply an individualization checklist before finalizing any goal.
- Flag any phrasing used verbatim across more than 5% of your caseload.
- Embed culturally responsive language referencing home language supports and community contexts.
Principle 4: Transparency and Family Communication
Core Idea: Families are essential partners who deserve clear information about AI use.
Action Steps:
- Add an "AI-Use Disclosure" statement to IEP documents.
- Provide plain-language guides explaining AI's role and limitations.
- Document family consent or concerns before finalizing AI-drafted content.
Principle 5: Collaborative Practice and Reflection
Core Idea: IEPs are inherently collaborative, and structured review prevents over-reliance on AI.
Action Steps:
- Share polished "discussion drafts" with team members at least 48 hours before meetings.
- Allocate meeting time to collaboratively review and adjust content rather than generating it in real-time.
- Conclude meetings with brief reflections on which AI-informed elements worked and how to refine future prompts.
Principle 6: Feedback, Monitoring, and Continuous Learning
Core Idea: Ethical AI use demands ongoing iteration and reflection.
Action Steps:
- Track metrics such as revision cycles, time allocation, and the frequency of bias or compliance issues.
- Compare AI-assisted and human-only drafts to understand patterns.
- Integrate quarterly reflections into professional learning communities and update practices based on findings.
Framework for Ethical AI Use in IEP Development
Practical Implementation: The Decision Tree
To support real-time decision-making, we've developed a decision tree that guides educators through key checkpoints:
Implications for Teacher Preparation and Professional Development
This framework has immediate applications for both in-service educators and teacher preparation programs. Each principle includes specific assignment examples for preservice teachers and application examples for practicing educators.
For example, TPPs might require candidates to:
- Draft goals both with and without AI, maintaining edit logs, and reflecting on compliance challenges.
- Exchange prompt templates with peers and apply bias check rubrics to each other's work.
- Develop plain-language parent guides explaining AI use with sample consent forms.
In-service educators might:
- Participate in monthly 15-minute PD sessions on prompt literacy.
- Conduct prompt portfolio reviews during PLC meetings.
- Track metrics over time to refine their AI-assisted workflows.
Moving Forward Responsibly
GenAI tools are here to stay, and when used responsibly, they can help alleviate the very real burdens special educators face while improving IEP equal access. However, these tools can never replace the human expertise, professional judgment, and commitment to individual students that lie at the heart of special education. The framework we propose doesn't offer a simple yes-or-no answer to AI use in IEP development. Instead, it provides a structured pathway for educators to engage critically and ethically with these powerful tools—harnessing their benefits while safeguarding student rights, educational equal access, and the dignity of every student. As Mr. Schrute learned, efficiency without attentiveness can compromise student outcomes. But with the right framework, training, and commitment to ethical practice, AI can become a valuable collaborator in our ongoing work to ensure a free and appropriate public education (FAPE) for all students with disabilities.
Read the full article: Coleman, O. F., & Waterfield, D. A. (2026). Ethical AI use in IEP development: A guiding framework. Journal of Special Education Technology, 0(0), 1-8. https://doi.org/10.1177/01626434261419099
References
eSchool News Staff. (2025, July 1). First AI-powered special education management platform helps districts reach 100% IEP compliance. eSchool News.
https://www.eschoolnews.com/newsline/2025/07/01/first-ai-powered-special-education-management-platform-helps-districts-reach100-iep-compliance/
Evans, R., & Sinha, N. (2024, May). Bridging the gap: Diversity initiatives in AI education. In Proceedings of the AAAI Symposium Series, 3(1), 474-477.
https://doi.org/10.1609/aaaiss.v3i1.31260
Dodge, J., Sap, M., Marasović, A., Agnew, W., Ilharco, G., Groeneveld, D., Mitchell, M., & Gardner, M. (2021). Documenting large webtext corpora: A case study on the colossal clean crawled corpus. arXiv preprint arXiv:2104.08758.
Glazko, K., Mohammed, Y., Kosa, B., Potluri, V., & Mankoff, J. (2024, June). Identifying and improving disability bias in GPT-based resume screening. In Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency (pp. 687-700). https://doi.org/10.1145/3630106.3658933
Imada, B. (2024, April 1). Generative AI’s impact on students of color and diverse students. USC Annenberg Relevance Report.
Jin, L., Shen, Z., Alhur, A. A., & Naeem, S. B. (2025). Exploring the determinants and effects of artificial intelligence (AI) hallucination exposure on generative AI adoption in healthcare. Information Development. https://doi.org/10.1177/02666669251340954
Kolisko, S., & Anderson, C. J. (2023). Exploring social biases of large language models in a college artificial intelligence course. In Proceedings of the AAAI Conference on Artificial Intelligence 37(13), 15825-15833. https://doi.org/10.1609/aaai.v37i13.26879
Nixon, N., Lin, Y., & Snow, L. (2024). Catalyzing equity in STEM teams: Harnessing generative
AI for inclusion and diversity. Policy Insights from the Behavioral and Brain Sciences, 11(1), 85-92. https://doi.org/10.1177/23727322231220356
Ocumpaugh, J., Roscoe, R. D., Baker, R. S., Hutt, S., & Aguilar, S. J. (2024). Toward asset-based instruction and assessment in artificial intelligence in education. International Journal of Artificial Intelligence in Education, 34(4), 1559-1598.
https://doi.org/10.1007/s40593-023-00382-x
Perez, F. P. (2024, September 25). Creating IEPs with GenAI while ensuring data privacy.
eSchool News. https://www.eschoolnews.com/it-leadership/2024/09/25/creating-ieps-with-genai-while-ensuring-data-privacy/
Rakap, S. (2024). Chatting with GPT: Enhancing individualized education program goal development for novice special education teachers. Journal of Special Education Technology, 39(3), 339-348. https://doi.org/10.1177/01626434231211295
U.S. Department of Education, Office of Educational Technology (2023). Artificial intelligence
and future of teaching and learning: Insights and recommendations. Washington, DC. https://www.ed.gov/sites/ed/files/documents/ai-report/ai-report.pdf
Waterfield, D. A., Coleman, O. F., Welker, N. P., Kennedy, M. J., McDonald, S. D., & Cook, B. (2025). IEPs in the age of AI: Examining IEP goals written with and without ChatGPT. Journal of Special Education Technology, 0(0). Advance online publication. https://doi.org/10.1177/01626434251324592
Join the conversation in our community!
CIDDL is committed to providing high-quality resources to support the growing knowledge, adoption, and use of a range of educational technologies for educators, related services, and leadership preparation programs. For more resources, including videos and blogs, subscribe to our newsletter and follow us on YouTube, Facebook, and LinkedIn. The most important part of our CIDDL community is YOU. Join our community and share the innovative ways you are using technology, ask a question about technology integration, or participate in our bi-weekly live AI Community Chats. We look forward to seeing you in our community!