How Artificial Intelligence Helps Blind People Read Ebooks – Today and Tomorrow

Only ten years ago, blind readers had access to little more than specialist software and a limited number of electronic books. Today, thanks to artificial intelligence (AI), there are many new possibilities. AI not only facilitates reading but also changes the way in which people with visual impairments can access literature. How does it work in practice? What is already working and what is still being developed?

 A New Generation of Speech Synthesisers

Traditional speech synthesisers like JAWS or NVDA have functioned in a “mechanical” form for years. AI has changed the landscape drastically. Natural-sounding voices – like those available in Amazon Polly, Google Cloud Text-to-Speech or Microsoft Azure services – were created thanks to machine learning. They can be integrated with ebook-reading apps, like Voice Dream Reader or Speech Central.

A Real-Life Example: today, it takes a blind person a few seconds to turn an EPUB or PDF file into a smoothly read ebook using a voice which resembles that of a real person. Some of these voices (e.g. Microsoft’s “Jenny Neutral”) are of almost radio quality, preserving the intonation, rhythm and emotions.

 Automatic Book Translations

Thanks to language models (e.g. ChatGPT or DeepL), blind individuals can read books in foreign languages with live support. You just have to copy a part of the text and paste it into the translating tool. More and more apps directly integrate these functions. For example, Readwise Reader enables quick translation of books or articles.

What is more, AI can “understand” the context – it doesn’t translate word for word anymore. Instead, it translates the meaning of a paragraph. This is particularly important for blind individuals who often use voice translators and don’t have visual indicators like footnotes, text formatting or graphs.

Intelligent Search and Book Navigation

Blind individuals often have to “listen to” the whole book in order to find the part they are interested in. AI changes this. Modern readers supported by language models can answer questions like “where in the book is the description of the Battle of Grunwald?” straight away and take the reader to the right place at once.

Solutions in Practice:

• Kindle + VoiceView (with the Alexa assistant): allows you to ask questions about a book’s content.

• BARD AI and ChatGPT: once a text file has been scanned you can ask questions and receive contextual answers.

 Automatic descriptions of graphs, illustrations and diagrams

Thanks to AI it is possible to describe graphics. If an ebook contains a graph, a vision language model (e.g. GPT-4o) can “see” the graph and describe it. It facilitates things greatly, as in the past such descriptions had to be prepared by humans.

Example: A user opens an economics textbook which contains graphs and illustrations – an AI-based app (e.g. Microsoft’s Seeing AI) can describe them and interpret data.

 Books Read Live by a Camera

AI supports OCR technologies (recognising text) – today, it is not necessary to scan a paper book for it to be read by apps like Envision AI, Seeing AI or Google Lookout. The book only needs to be “shown” to a phone camera. AI recognises text, language and context and transforms it into speech. It can also translate in real time.

 Context Assistants and Book Summaries

Thanks to AI, you can not only read but also understand books. Models like Claude, Gemini or ChatGPT can summarise a chapter, explain a difficult subject and even discuss a novel’s main topics.

Who benefits from it? It enables blind individuals who study or are preparing for exams to learn faster without having to listen to the whole publication.

What might appear in the foreseeable future?

• Interactive ebooks with AI assistants – books which talk to the reader and answer questions.

• AI personalising the speed and way of reading – recognising when the reader looses track and automatically slowing down or repeating content.

• Readers with cameras recognising emotions and attention levels – which adjust the narration style.

Summary

Artificial intelligence is no longer a technological peculiarity and becomes a real support tool. For blind and visually impaired individuals it means more independence as well as a deeper, more personal way of experiencing literature. AI makes it possible to read faster, more easily and more consciously – which is what reading is all about.

The article was created in close collaboration between the Polish Foundation for the Blind and Visually Impaired "Trakt” and Have a Book.

Translated by Aleksandra Kallas