How to write alt text with AI as a starting point
Many people assume describing an image for the web requires listing every single visual detail. This assumption leads to long, cluttered descriptions that distract from the main message of the page. When content creators learn how to write alt text, they discover that the task is actually about conveying the meaning of an image within its specific context.
The purpose of an image on a page
An image on a website is rarely just a decoration. It usually serves a specific purpose, such as explaining a concept, showing a relationship, or setting a mood. The alternative text must reflect this purpose rather than just listing what is in the picture.
If an image contains a chart showing sales growth, the description should not just say that there is a line graph with a blue line. It should state the main trend that the chart is there to show. The reader needs the information that the chart conveys, not just a description of the lines and colours.
When the image is purely decorative and does not add information, it needs empty alternative text. This tells screen readers to skip the image entirely, which keeps the reading experience clean and focused.
The hidden cost of manual descriptions
Writing these descriptions manually for every single image on a website takes a significant amount of time. It requires looking at each image, deciding its purpose, and drafting a description from scratch.
This process often leads to delays in publishing content. Because it is time-consuming, some teams skip it entirely, which makes their websites inaccessible to many users. The cost is not just the time spent drafting, but the potential loss of readers who cannot access the information.
When teams try to describe every detail, the descriptions become too long. This makes the page harder to navigate for people who use screen readers, as they have to listen to long descriptions that do not help them understand the page.
The draft from the machine
An artificial intelligence tool can now suggest a description for an image. It looks at the visual elements and provides a starting point for the text.
This suggestion is not the final version, but it provides a draft that the author can work with. The machine handles the initial translation of visual elements into words, which saves the author from starting with a blank page.
By using an AI tool to generate these drafts, the time required to create alternative text is greatly reduced. The tool can process many images quickly, providing a consistent starting point for the entire website.
The author's final decision
The author keeps the most important part of the task, which is deciding if the suggested description fits the purpose of the page. A machine can describe what is in an image, but it does not know why the author chose to put that image on that specific page.
The person reviews the suggestion and edits it to match the context. They make sure the description is short and direct, as long descriptions should be avoided in alternative text.
If the image contains complex information, like a detailed chart, the person decides to put that long information in the main page text instead. They ensure the alternative text remains brief, while the detailed analysis is accessible to everyone in the body of the article.
Different ways of working
Using AI as a starting point changes how teams approach accessibility. It moves the task from a tedious chore at the end of a project to a standard part of the publishing workflow.
This approach can be applied to other areas of content creation, such as making a staff training video from a policy. In both cases, the machine drafts the initial content based on existing material, and the person refines it for the audience.
To make this process work smoothly, teams can set your AIs rules once with custom instructions to make AI accessible. This ensures the tool suggests descriptions that already follow the correct guidelines, reducing the amount of editing needed.
The path to accessible content
When the repetitive drafting is handled by a machine, people can focus on the decisions that require human judgment. They can spend their time making sure the content is clear, accurate, and helpful for all readers.
This shift allows more people to participate in creating accessible content. It removes the barrier of writing descriptions from scratch, making it easier for everyone to contribute to a more accessible web.
By focusing on the purpose of the image rather than its details, authors create a better experience for their readers.
This collaborative approach to drafting makes digital spaces open to everyone.
If you want your own website, app or documents to work for everyone, see digital accessibility at aXai: audits and fixes to WCAG, and Easy Read versions of what you already publish.
aXai is a disability-led studio that turns ideas into software and content with AI. AI takes the glue work — the chasing, the forms, the fourteenth retelling. People keep the part that was always theirs: deciding what is worth doing, and making it. New ideas. A different way of working. Read how AI-supported work works.
Easy Read version
What people think
Many people think alt text must describe every detail.
This is not true.
Alt text should tell you what the image means on that page.
How to do it
Keep your alt text short.
If an image is just for decoration, leave the alt text empty.
If an image has a chart, put the details in the main text.
How AI helps
An AI tool can write a draft description for you.
This saves you time.
You do not have to start from a blank page.
The human part
You must check the draft.
You decide if the description fits the page.
Only a person can decide what the image really means.
aXai helps make websites and apps work for everyone.
See digital accessibility at aXai.
aXai is a studio led by people with disability.
AI does the chasing and the forms.
People decide what is worth doing.
People make it.
That is a different way of working.
