AI, inclusion and students with disability

Your staff are already using AI on student data.

Dr K. Clark Burt — Inclusive education & educational technology

Disability Inclusion documentation is a significant administrative load, and generative AI looks like the obvious answer. Departmental policy restricts putting personally identifiable information into these tools. Most teachers doing it don't know that.

I'll show your staff where AI genuinely helps students with disability, where it exposes your school, and how to tell the difference.

PhD, University of Melbourne · Inclusive Education resources for the Digital Technologies Hub, Australian Government Department of Education · Digital Technologies curriculum levels A–D, Victorian Department of Education

Student support plan Adjustments record
Student Aisha M. Rahman
Year level Year 8
Diagnosis Mild intellectual disability; expressive language disorder
Funding tier Tier 3
Carer contact S. Rahman — 04XX XXX XXX
Adjustments Reduced text density; visual task cards; extra processing time
Review notes Ongoing difficulty with multi-step written instructions

Everything blacked out here is personally identifiable information. Departmental policy restricts putting it into a generative AI tool.

Do not copy/paste

Work has included

  • Education Services Australia Inclusive Education resources for the Digital Technologies Hub
  • Victorian Department of Education Digital Technologies curriculum, levels A–D
  • University of Melbourne PhD; lecturer and tutor
  • QUT Online Lecturer
  • Digital Learning & Teaching Victoria Past President
  • Australian Association of Special Education Victorian Chapter

The tools are useful. The line isn't where people think it is.

Victorian schools have real obligations under the Disability Standards for Education and real documentation to produce under Disability Inclusion. The workload is genuine and teachers are looking for relief.

Departmental policy restricts uploading personally identifiable information into generative AI tools. A student's profile — name, functional needs, adjustments, often health information — is about as identifiable as school data gets. Nobody set out to breach anything. The policy is buried and the deadline isn't.

Student information going into public tools

Documentation drafted with AI assistance, with the student's details still in the prompt. Fast, effective, and outside policy.

Resources adapted without understanding why they failed

AI will happily simplify a text. Whether the result matches how the student actually processes language is a separate question, and the tool can't answer it.

Nobody who can say where the line is

Staff are making individual judgement calls about privacy, disability and technology, without a shared position or anyone qualified to set one.

Three ways to work together

AI, safely, with students with disability

A day with your staff on what these tools can and can't do for students with disability, and where your obligations sit.

  • Where it genuinely helps: adapting resources, reducing language load, building visual supports — with no student data involved
  • Where it exposes you: documentation, profiles, anything with a name attached
  • What departmental policy actually requires, in plain language
  • Practical work on your own resources, not generic examples

One day, on site around Melbourne or online · $960

Resource review

You send me the resource. I tell you what will fail and why.

  • A written report on where the resource is too language-intense, visually cluttered, or unworkable in a mixed classroom
  • Specific changes, section by section, grounded in cognitive load theory
  • Notes on accommodations for students with disability learning alongside neurotypical peers

Typically 3–5 days · $2,900–$4,800

Resource redevelopment

The review, plus the changes made.

  • Rewritten text at an appropriate language load
  • Restructured layout, increased visual support, clutter removed
  • Returned ready to publish

Scope and price quoted after an initial review

For Victorian government schools: Disability Inclusion Tier 2 school-level funding exists to strengthen school capability in inclusive practice. That's the line the professional learning day usually comes out of.

How it works

A conversation, no charge

Twenty minutes. What you're using, what's worrying you, whether I'm the right person.

I look at what you actually have

Your resources, your documentation process, what staff are really doing. Not what the policy says should be happening.

You get specifics

Not "consider reviewing your AI practices." Which practice, why it's a problem, what to do instead.

You decide what happens next

A PD day, a resource review, or nothing. Sometimes the answer is that you're fine and don't need me.

I'll tell you what's wrong. That's the job. Most people who hire me have already had someone tell them it looks fine.

Case study

Generative AI in an accredited VET course, with students with intellectual disability

Year 11 and 12 students from a specialist school undertook a Certificate II in Digital Technologies at Holmesglen TAFE as part of a pilot program. I co-taught the course with the TAFE teacher.

One unit asked students to produce a podcast. They worked with software none of them had used before — audio editing tools and generative AI — building an episode on a topic they chose themselves.

No student information went into any tool. The AI was used on the work, not on the students.

The conventional expectation is that these students need the technology simplified. They didn't. They needed the language load reduced and the task made clear. Given that, they used unfamiliar professional software to make something they cared about, inside an accredited qualification.

That's the whole argument. Used properly, these tools raise the ceiling. The ceiling is usually in the resource, not the student.

Why me

PhD, University of Melbourne

Computer-assisted instruction for teaching vocabulary to students with mild intellectual disability. I built the web server, database and application that ran the study myself.

Education Services Australia

Part of the team that developed the Inclusive Education resources for the Digital Technologies Hub, published by the Australian Government Department of Education. Engaged again afterwards to review and advise on resources.

Victorian Department of Education

Contributed to the Digital Technologies curriculum for levels A–D.

The theory behind VTLM 2.0 is the theory behind my doctorate

VTLM 2.0 has explicit teaching at its core, and explicit teaching rests on managing cognitive load. That's what I spent my PhD on, with the students for whom it isn't optional.

Digital Learning & Teaching Victoria

Past President, 2024–25. Six years on the committee.

Still teaching

Since 2004, 18 of those years in Australia, and at a specialist school for students with mild intellectual disability since 2013. I'm not advising you from theory.

Clark Burt, holding a mug reading “Keep calm and pretend it’s on the lesson plan”.

About

I'm an educational technologies consultant specialising in inclusive and special education.

I started as a computer programmer and have been teaching since 2004 — first in the United States, then in Australia from 2008. My doctorate at the University of Melbourne looked at how students with mild intellectual disability learn vocabulary from computer-assisted instruction. I built the system that ran the study myself, which is roughly how I've always worked: the technology and the teaching aren't separate problems.

I currently teach digital technologies at a specialist school, lecture in educational research at QUT Online, and do relief teaching across Victorian schools. It means I see what's actually happening in classrooms rather than what's reported to be happening.

I'm straightforward about what I find. Clients don't always enjoy it. It's what they're paying for.

Questions

What does it cost?

A professional learning day is $960. A resource review is typically $2,900–$4,800 depending on size and complexity. Redevelopment is quoted after a review. Day rate is $960.

Should we host our own AI model so data stays in the school?

Usually no. Keeping data on site solves one problem — it doesn't remove your obligations around consent, accuracy, retention or accountability for what the model produces, and sensitive information stays sensitive wherever it's processed.

There's also the question of who maintains it, monitors output quality, and is responsible when it writes something wrong into a student's record. Most schools don't have that capacity. I'd rather help you work out what you actually need before anyone buys hardware.

Are you telling us not to use AI?

No. I use these tools with students myself. The problem isn't the technology, it's using it on the wrong things without knowing it.

We've already got an AI policy. Isn't that covered?

A policy tells staff what they should do. It doesn't tell you what they are doing, and it rarely addresses students with disability specifically — which is where the data is most sensitive and the stakes are highest.

Does this only apply to students with disability?

The privacy exposure applies to every student. The design work applies more widely than people expect: resources built for a narrower working memory for language are easier for everyone. The disability case just makes the failure visible.

We're not in Melbourne. Can you still help?

Reviews and advisory work are done remotely. On-site professional learning is around Melbourne and includes one-way travel time in the quote.

Are you a lawyer or a privacy consultant?

No. I can tell you what departmental policy says, what it means in a classroom, and where I'd be careful. Formal legal or privacy advice you should get from someone qualified to give it.

Start with a conversation

Tell me what your staff are using and what's making you uneasy about it. If it's not something I can help with, I'll say so.

No obligation. I'll reply within two business days.