Beyond basic speech to text capacities, what can AI Reading and Accessibility Tools do and what are some issues to consider with them?

‍

Over the last few years, reading and accessibility tools have quietly evolved from simple text‑to‑speech apps into more complex, AI‑assisted ecosystems. Products such as Read&Write, Immersive Reader, and similar browser extensions can sit on top of web pages, PDFs, and Google Docs. In doing so, they can offer features such as read‑aloud, line‑by‑line highlighting, picture dictionaries, vocabulary lists, and word prediction. The aim of these kinds of tools is aim is to make text more accessible for:

‍

1. Dyslexic learners
2. Students reading in a second or third language
3. Anyone who may find dense pages visually or cognitively overwhelming.​
4. ‍

Recently, a new layer has started appearing: tools which don’t just read text, but also can also transform it. For example, an AI can simplify a passage, generate a summary, or rephrase a paragraph at a different reading level. Some platforms can even create quizzes, glossaries, and comprehension questions automatically, based on the text in front of the learner. For younger children, apps such as Ello AI Reading Coach listen as the child reads aloud, using speech recognition and machine learning to track fluency, flag mispronunciations, and offer encouragement or corrections in real time.​​

‍

Crucially, these tools are no longer always niche special‑education software; they are increasingly bundled into mainstream platforms and available as free or low‑cost browser add‑ons. This means families can try them at home without formal diagnoses, and teachers can layer them into everyday lessons rather than reserving them for a small group. 

‍

However, it also raises questions about over‑reliance, data collection, and what “independent reading” should mean in a world where almost any text can now be modified or read aloud on demand.​​

Where These Tools Can Help & Issues to Keep in Mind

At their best, AI‑powered reading and accessibility tools should support, rather than act for. They may allow students to access age‑appropriate ideas, stories, and information where the form may otherwise put said ideas, stories and information beyond reach. For example, for a dyslexic 12‑year‑old fascinated by astronomy, having a complex article read aloud (with highlighting of the text following the voice) could be the difference between frustration and genuine intellectual engagement. For multilingual families, in‑line translation and picture dictionaries could turn English‑only materials into something more navigable without having to wait for materials to be rewritten by a teacher.​

‍

From a position of empowering, these tools could also be powerful tools for differentiation. For instance, a teacher could assign an article, then have students choose whether to read the original, a simplified version, or a read‑aloud depending on where they feel they’re at that day. At home, parents could gradually fade support by shifting from full‑text read‑aloud to selective use (for tricky paragraphs only) as decoding and comprehension skills improved. Learners with ADHD could use features such as line focus, background colour changes, and on‑the‑fly audio to reduce visual clutter and cognitive load.​​

‍

The risks start when this support becomes the only road and hobbles rather than supports growth and confidence. If every text were auto‑simplified and always read aloud, students might never practice the effortful but necessary skills of decoding, inferring meaning from context, and grappling with complex syntax and sentence structure. There is value for example in being able to read Shakespeare or T. S. Elliot in its original text. The message can be in the medium, and this is also often where our culture sits. Simplification algorithms can also distort nuance – hedging language, irony, or technical precision may be flattened into something more bland or even misleading. 

‍

Regarding data and privacy, detailed logs of what a child reads, where they click and after how long, and how they pronounce words can be extremely sensitive, especially when tied to names, emails, or school accounts.​

‍

Keeping the above in mind, the key might be not to reject these tools, but to use them with caution and intention. Families and schools can be explicit about when a tool is being used to access content (fair and necessary) versus when the goal is skill‑building (where some desirable difficulty is important). Regularly asking, “What part of this reading should feel a bit challenging for you right now?” keeps the human learner – not the software – in charge of their own learning journey.

‍

Note: the first draft of this article was done by AI Chatbot Claude with the support of Max Capacity. The text was then edited and adapted by Jaye Sergeant of Turtle & elephant, who is responsible for the published version.

‍