AI can help with IEP documentation in ways that save real time. It can also produce confident-sounding nonsense that will cause problems. This is a clear-eyed guide to what AI should and should not touch in the IEP process.
The Individual Education Plan is one of the most important documents in a student's school life. It is also one of the most time-consuming to write well. So it makes sense that educators are turning to AI tools to help. What is less clear is where the line is between genuinely useful AI assistance and AI assistance that creates legal and ethical problems.
This is my attempt at a clear, practical answer to that question, based on real experience with both IEPs and AI tools.
What AI Can Do Well
AI is genuinely good at the drafting and language work that makes IEP writing time-consuming without making it substantively better. Specifically:
- Drafting SMART goal language when you give it the area of need and current performance level
- Generating benchmark milestones to sit under a broader annual goal
- Rewriting vague goal language into specific, measurable terms
- Suggesting accommodations and adjustments by area of need
- Converting clinical language from assessment reports into plain terms for the 'current level of performance' section
- Formatting and structuring documentation so it meets expected conventions
- Drafting parent-friendly summaries of IEP content
What AI Cannot Do
AI does not know the student. This sounds obvious but it has real implications. An AI tool drafting IEP goals does not know that this particular student becomes completely dysregulated during fire drills and that the transition plan needs to account for that. It does not know that this student's relationship with their ESW is the primary reason they attend school consistently. It does not know the family's circumstances, communication preferences, or the history of this plan.
- Determining the student's actual area of need (that requires assessment and relationship)
- Deciding what goals are appropriate given the student's history and trajectory
- Making legal determinations about what adjustments are required
- Replacing the planning meeting and the voice of the student and family
- Knowing whether a suggested accommodation will actually work for this student
- Assessing current performance levels (requires direct observation and formal assessment)
“An IEP is a legal document describing the real needs of a real person. AI can help you write it better and faster. It cannot substitute for knowing the student.”
Where to Draw the Line
The line is clearer than people sometimes make it. AI is appropriate where the task is drafting language, formatting, and structuring information you already have. AI is not appropriate where the task is determining what that information should be.
If you already know that the student's goal is to improve their ability to initiate peer interactions in unstructured settings, AI can help you write that goal well. AI should not be deciding what the goal should be.
Never let AI generate the 'current level of performance' section from scratch. This section must reflect real assessment data and real observation. AI-generated performance descriptions that have not been grounded in actual data are a professional and legal risk.
Privacy and Data
Do not paste identifiable student information into public AI tools. This means no full names, dates of birth, school names, or information that could identify a specific student. Use initials or placeholder names, describe situations in general terms, and be aware of your school's data privacy policies. Some AI platforms designed for educators have specific data handling agreements, but free consumer tools like ChatGPT do not offer the same protections.
A Practical Workflow
- Gather your assessment data, observation notes, and prior plan as you normally would
- Use AI to help you draft goal language from the areas of need you have already identified
- Review every suggestion against what you actually know about the student
- Delete anything that is not accurate or does not fit, no matter how well it is written
- Do not let polished language substitute for accurate content
- Ensure the final document reflects the real student and is defensible if questioned
The Bottom Line
AI is a drafting tool for IEPs, not a planning tool. Used with that understanding, it can genuinely reduce the time burden of IEP documentation without compromising the quality or accuracy of what you produce. Used without that understanding, it creates the risk of polished, confident documentation that does not reflect the actual student, which is worse than documentation that is awkward but accurate.
A note on accuracy:While every effort has been made to ensure the information in this article is accurate at the time of writing, facts, policies and research can change. We're human, and sometimes we get things wrong. If you spot something that needs updating, we'd genuinely love to hear from you.
Newsletter
Worth reading. Not often.
Practical guides on neurodiversity, NDIS navigation, and Australian schools. Sent when there's something worth saying, not on a schedule for the sake of it.
No spam. Unsubscribe any time.

Dave Harrison
ESW · Neurodiversity Advocate · Podcast Host
Dave Harrison is currently working in Australian schools as an Education Support Worker. He's the founder of THRVHUB, host of the Different Is Normal podcast, and a parent of a neurodivergent teenager, writing from both sides of the classroom.
More about DaveMore to read
Comments
Sign in with GitHub to leave a comment. All comments are moderated through GitHub Discussions: respectful and on-topic only.