Picture this: you’ve created a detailed diagram that explains a complex topic. Can all your students access that visual information? For disabled people who use screen readers, or those viewing content with images turned off, these visuals become invisible barriers unless you provide alternative text.
Alternative text, often called “alt text”, describes images for people who cannot see them. Writing effective descriptions traditionally requires time, skill, and subject expertise. What if artificial intelligence could help you create better alt text faster?
Why alternative text matters
Beyond supporting disabled students, alt text benefits everyone.
- Students listening to audio versions of materials rely on these descriptions.
- Those on limited data plans who disable images won’t miss crucial information.
- International students gain additional context when processing content.
- Neurodivergent students may appreciate comparing the written description with the visual artefact.
However, when workloads have never been higher, writing alt text can feel overwhelming. Here’s where AI transforms the task from burden to opportunity.
From overwhelming to achievable
Tools like Microsoft Copilot can examine charts, photographs, diagrams, and illustrations, producing both concise alt text and comprehensive descriptions.
The magic happens when you combine AI capability with your subject expertise. The AI handles the initial heavy lifting, identifying colours, shapes, text, and relationships within the image. You refine the output, ensuring it captures the educational purpose and context that only you understand.
Consider a graph showing climate temperature changes over time. AI might describe: “Chart with blue and red sections showing temperature data overtime”. Your refinement adds educational significance: “Chart showing global temperature anomalies from 1850 to 2020, demonstrating clear warming trend since 1980.”
This approach amplifies your expertise whilst dramatically reducing time investment.
How much alt text should we write?
- If we’re just using the image for visual interest, we’ll mark it as decorative.
- If the image replicates text already in the resource in which we will use the image, then we should add a caption describing the essential content and context and use the alt text feature to mark the image as decorative.
- If the image supports or rationalizes what is in the text, then we should write a brief alt text describing essentials of the image and use a caption or title.
- If the image goes further than what is in the text, or I want students to use the image in some way, then I should write a brief alternative text describing the essentials, add a caption, and then provide a full description elsewhere in the document.
- Ideally, we should also provide any tabular data that was used to generate the image, for example if it was a chart or graph.
Bite-sized task
Step 1 – learn
Open Microsoft Copilot using your University login and do the following:
Upload an appropriate image that you use in teaching and ask Copilot to describe it.
Using this basic instruction with no extra guidance shows you what Copilot can do without context.
Step 2 – do
Next time, use the following instructions with the same image.
Before you start working with Copilot, take a moment to think about the image’s context of use.
Think about:
- Subject area: What discipline or field?
- Course level: Undergraduate year 1, 2, 3, or postgraduate?
- Specific topic: What concept, theory, or skill does this lesson cover?
- Learning objective: What should students understand after engaging with this image?
Your Task When you have both the image and context, write an instruction for CoPilot that begins ‘Create Alt text for this image. Here is some context…[then give the context for the image i.e. include the subject area, course level, specific topic and learning objective].’
Optional: How specific do you want to be? : There may be times when it is helpful to create extended descriptions of images (for example to describe a concept in your learning resource). You can use the some of the prompts below to create highly detailed descriptions, but think about whether this adds value in your specific context to justify the additional effort…
1. Title Write a brief, descriptive title for the image that reflects its educational purpose.
2. Alt Text (Two Sentences Maximum)
- First sentence: Describe what the image shows (the facts, construction, visual elements).
- Second sentence: State the educational significance within the given course context.
3. Long Description Provide a detailed description organised with:
- Clear headings
- Paragraph breaks
- Bullet points where appropriate
- Complete coverage of all visual elements
- Educational significance: Explain how this image supports the specific learning objectives you were given.
This description must enable someone who cannot see the image to understand both what it contains and how it advances their learning in this particular course.
4. Text Content List all words, numbers, and text that appear in the image exactly as written.
Writing Requirements
- Use active voice throughout.
- Address the reader directly Use British English spelling (colour, centre, organised, analyse).
- Write numbers one through nine as words, 10+ as digits.
- Keep sentences concise and clear.
- Connect to course context: Always link visual elements to specific learning objectives.
- Consider how the image fits within the broader curriculum.
- Explain why this particular visualisation matters for these students at this level.
Compare the resulting pedagogically-focused descriptions with the generic description from your first prompt. At each point, how has AI performed?
Step 3 – reflect
Consider these questions and share your thoughts on the weekly Teams post:
- How accurately did copilot identify the key elements in your image? What important details did it miss or misinterpret?
- How did the initial description differ from the second description? What subject-specific knowledge did you add?
- How might this workflow change your approach to creating visual content? Would you design images differently knowing Copilot can help describe them?
- What concerns do you have about relying on Copilot for accessibility tasks?
Join the conversation
Post your thoughts on the weekly Teams post to join the conversation.
Further links
Find out more about the Blackboard Ally accessibility tool.
The Introduction to Digital Accessibility course on Blackboard is open to all staff and students.
To learn more about alternative text, watch the recorded webinar, What is alternative text? How do I write it for images, charts, and graphs? Or read the transcript.
For guidance on writing alternative text for highly complex images, review Complex images – making sense for accessibility.
Contributor biography
Matthew Deeprose is based in iSolutions. He has given numerous presentations on accessibility and AI to international audiences, his presentations, blog posts and accessibility tools are available on his website. He was co-chair of the International Ally User Group from 2022 to 2025. He has won two Blackboard Catalyst Awards, The JISC Community Champion 2022 award and was a finalist in the EdTech Awards 2022. He is a member of the UCISA Diversity Taskforce, the University of Southampton Digital Equality Steering Group and the iSolutions EDI group.
© 2025. This work is openly licensed via CC BY-NC-SA