AI for SEND and inclusive teaching: practical applications

Of all the uses for AI in further education, support for learners with SEND is one of the least controversial and, for many staff, one of the most immediately useful. It is worth treating this as its own topic rather than a footnote to general AI use, because the applications and the risks are genuinely different.

Text to speech and reading support

AI-powered text to speech tools have improved considerably, moving well beyond the robotic voices of a few years ago to natural-sounding narration that learners are more willing to actually use. For learners with dyslexia, visual impairment or processing difficulties, this can mean independent access to written materials that previously required a support worker reading alongside them in every session. Many tools now also offer synchronised text highlighting as the audio plays, which supports both listening and reading simultaneously, a technique with a solid track record in literacy support.

Reading level adaptation

A single handout can now be produced at multiple reading levels quickly: the same content, the same key vocabulary introduced deliberately, but simplified sentence structure and shorter paragraphs for learners who need it. This is particularly useful in functional skills and Entry Level provision, where the same topic often needs to serve learners at genuinely different reading ages within one group. The key discipline here is checking that simplification has not accidentally removed or distorted the actual content, since AI tools sometimes simplify by cutting detail rather than rephrasing it.

Translation for ESOL and EAL learners

Providing materials in a learner's first language alongside English supports comprehension without removing the English language learning goal, since materials can be presented bilingually rather than as a replacement. AI translation has become good enough for general classroom materials, though it remains unreliable for technical or specialist vocabulary, so translated versions of safety-critical or assessment materials still need a human check, ideally from a fluent speaker.

Alternative formats

Learners with different needs benefit from the same content presented differently: a written explanation converted into a simple diagram description, a long document turned into a bullet-point summary, or a text converted into a suggested structure for an audio recording rather than written submission where a qualification allows it. AI tools make producing these alternatives fast enough that it becomes practical to offer as standard, rather than as a special, resource-heavy accommodation arranged only after a request.

Communication support

For learners with speech, language and communication needs, AI-assisted communication aids, including predictive text and symbol-based communication tools that use AI to suggest phrasing, are increasingly capable and integrated into everyday devices rather than requiring separate specialist hardware.

Supporting processing and organisation

Some learners with ADHD or executive function difficulties benefit from AI tools that turn a task description into a step-by-step checklist, or a long piece of written work into a suggested structure before they start writing. This scaffolding support can reduce the barrier to starting a task, which is often the hardest part for these learners.

Where caution is genuinely needed

• Do not rely on AI-generated content as a full substitute for a specialist assessment of a learner's needs. AI can support delivery of an agreed strategy, it should not be setting the strategy itself.
• Be careful with data protection when describing a learner's needs to an AI tool in order to get tailored suggestions. Avoid naming the learner or including identifying detail, describe the need in general terms instead.
• Check that adapted materials still meet the actual assessment criteria, particularly when simplification for reading level touches on content a learner will be assessed against.
• Remember that some learners will find synthetic voices or AI-generated content harder to engage with, not easier; these tools are not universally beneficial and need to be offered as an option, not a default.

Building this into everyday practice

The most effective use of AI for inclusion is not a special separate process run by a learning support team alone, it is built into how every tutor prepares materials day to day, alongside the broader classroom uses covered in How teachers are using AI in the classroom: 12 real examples. When producing a differentiated worksheet or resource for a group with a learning support learner in it becomes a five-minute task rather than a fifty-minute one, inclusive practice becomes something that happens by default rather than only when time allows.

Training staff to use these tools well

Knowing that a tool exists is different from knowing how to use it well for a specific learner's actual needs. Staff need training on both the practical use of these tools and on the data protection and quality checks that keep their use safe, which fits well within broader digital literacy and inclusive practice training available through our qualifications and CPD courses.

A genuinely positive use case, with limits

Of all the AI applications covered across this cluster, inclusive teaching and SEND support is the one where the case for use is clearest and the ethical debate is quietest. That does not mean it is risk-free, but it does mean providers can move with more confidence here than in assessment-adjacent areas, provided the same basic discipline applies: check the output, protect learner data, and treat AI as a tool that supports a professional's judgement about an individual learner rather than one that replaces it.