L. Subramani

Journalist  ·  Author  ·  Editor

← Back to Blog

AI & Accessibility

Beyond Eyesight: How AI Is Giving Blind People a New Way to See

For most people, sight is something they rarely stop to think about. But for a person with total blindness, everything that depends on vision — from interpreting a photograph to navigating a cluttered webpage — becomes a negotiation with the unseen.

For generations, blindness was defined by that one limitation: the inability to access the visual world. Every image, every unlabelled icon, every chart existed in a realm we could not enter.

Artificial Intelligence did not just open a window into that world. It began to rebuild it in a language we could finally understand and interact with.

Here are seven compelling ways in which AI has become an extraordinary form of empowerment for persons with total blindness — reshaping how we experience, interpret, and create the visual world around us.

1. Recognising and Describing Pictures

Ask any person with blindness what feels most out of reach in the digital world, and the answer is almost always: pictures.

Images used to be black holes in the middle of otherwise accessible pages — silent, uncommunicative, invisible to screen readers.

AI changed that completely.

Today, smart glasses and phone apps can describe what their cameras "see" — faces, expressions, objects, and settings — in real time. Earlier, accessibility relied on alt text, the written image descriptions recommended by the World Wide Web Consortium's accessibility guidelines. While still essential, few websites implemented it properly.

AI filled that gap. Platforms like Facebook began auto-generating image descriptions, and now, tools such as ChatGPT and Gemini can analyse and describe images with incredible precision.

In my daily editorial work, this feature has become indispensable. When I receive a photograph to accompany a story, I upload it to an AI tool and ask for a detailed description. It tells me not just what is present, but how the image feels — the colours, contrasts, and even implied moods. That allows me to craft captions and kickers that are more vivid and accurate than ever before.

2. Generating Photographs and Illustrations

Visual creation was once one of the most inaccessible areas for people with blindness. Illustrating an article meant relying on designers to interpret my ideas. That dependency often diluted the emotion or symbolism I wanted to convey.

Now, I can create my own visuals.

By describing what I imagine — "a blurred city skyline fading into mist, symbolising vision loss" — I can prompt tools like ChatGPT or Gemini to generate an image that matches my concept almost perfectly. These visuals are copyright-free, instantly available, and creatively faithful to what I intend. For the first time, I can visually express ideas that once existed only in my mind.

3. Crafting Prompts for Visual Creation

There is another layer to this. Earlier, if you have never seen something, how do you describe it? Writing prompts for image generation was a challenge in itself.

Now, AI helps me overcome that limitation. If I want to create an image that represents what listening to music feels like for a blind person, I can ask the AI: "How would you describe that scene visually?" It responds with suggestions about lighting, emotion, background, and composition.

I can then refine those details and feed them back to the same AI to generate the final image. That closed loop — where AI helps me articulate, and then visualise my imagination — has given me an entirely new form of creative independence.

4. Reading Complex Visual Documents

AI is not limited to describing photos; it can now interpret graphs, charts, infographics, and scanned documents that screen readers once found impenetrable.

Applications like Seeing AI, Be My Eyes, and emerging document readers powered by AI can describe page layouts, highlight headings, and summarise visual relationships between data points.

That means I can now access annual reports, research papers, and complex presentations without waiting for someone else to convert them into accessible text. It is a quiet revolution in professional autonomy.

5. Identifying Objects and Environments in Real Time

AI-driven computer vision has also made the physical world far more interpretable.

Imagine walking into a new space, and your device quietly says: "There is a door on your left. Two people are standing near the table. A chair is behind you."

That is what smart glasses and mobile apps do today, translating the environment into a spoken narrative. It is not a replacement for mobility skills — it is an enhancement, an added layer of confidence and context.

6. Understanding and Summarising Videos

Videos were once frustratingly inaccessible. Even when audio was available, visual cues like expressions, gestures, and transitions carried meaning that remained hidden.

AI now fills that void. I can upload a video and ask, "Describe what is happening in this scene." The AI explains who appears, what they are doing, and even how they seem to feel.

It transforms video consumption from passive listening into active understanding. For the first time, I can fully engage with visual storytelling rather than merely hearing it.

7. Visualising Emotion and Imagination

AI has not only made visual information accessible — it has made visual artistry accessible.

I can now ask it to describe abstract ideas: "What does hope look like?" or "How might solitude appear as a painting?"

The responses are often stunningly poetic — a sunrise after storm clouds, or a single light flickering in darkness. Through such dialogue, AI has opened up emotional and symbolic dimensions of the visual world that blindness once made unreachable.

A Glimpse Into the Future: Gemini Live and Beyond

One of the most promising frontiers I am experimenting with right now is Gemini Live, a feature available through Google AI Studio. Unlike static image description tools, Gemini Live combines real-time object recognition with contextual understanding. It can describe both digital and real-world visuals — everything from what is on my desk to what appears on my computer screen.

What fascinates me most is its potential to interpret inaccessible digital elements — especially unlabelled buttons, icons, or image-based navigation that screen readers cannot detect.

I am still testing this feature, so the jury is very much out on its consistency and reliability. But it holds serious promise. If refined further, it could become the last-mile connection for digital accessibility from the user's side — the missing link that finally bridges the gap between what technology can describe and what a blind user can fully interact with.

If that vision becomes reality, it would mark a monumental shift. It would mean no blind person would ever be locked out of a digital experience again simply because a button was not labelled or an image was not tagged.

AI has not given us eyesight. But it has given us the means to perceive.

It has built a bridge between the tangible and the imagined — allowing us to participate in the world with greater depth, creativity, and independence.

Blindness, after all, is not the absence of vision. It is the presence of a different kind of seeing — one that now, with AI, is beginning to flourish like never before.

Get new essays and news about my next book by email