Improving Accessibility

A core competency for learning designers is to integrate inclusive design principles and practices in their work. One aspect of inclusive design relates to accessibility of digital assets, including documents, visual representations, and other types of media.

One aspect of digital accessibility is writing alternative text or alt text to provide a textual description of visual content for learners with visual impairments. Within some learning experiences, some visual elements might be decorative, while others could convey meaning that is relevant to understanding the content.

So, it is important for learning designers to be able to write alternative text to support comprehension of visual content elements that imbue meaning and not just decoration. Alt text can be embedded within various document types and is detectable by screen readers.

Writing alternative text is part art and part science. It should capture the meaning and essence of the visual element without being excessively descriptive. It should be accurate and also convey the intent of the image.

This key learning design task takes practice to develop proficiency, and even once proficient, it can be time-consuming to do so.

With that background in mind, let's consider how we can engage generative AI in the drafting of alternative text or alt text.

Let’s use an example from our course on extended reality in education. We can upload a slide to ChatGPT that includes icons associated with each of the learning affordances discussed throughout the course: spatial knowledge, experiential learning, intrinsic motivation, and knowledge transfer. And under each icon, there is a brief caption.

Hello ChatGPT, I have attached a slide showing the learning affordances of extended reality in education. Above each learning affordance is an icon. Please generate alternative text (alt text) for each of these icons.

Now, let's consider the output for each icon. ChatGPT has provided a description of the icon and an explanation of what each icon symbolizes. It's impressive to see that ChatGPT accurately described each icon and provided a reasonable explanation of the meaning ascribed to each one.

For example, intrinsic motivation—an icon that consists of a thumbs-up gesture representing approval, encouragement, and the motivational aspect of learning. The description provided by ChatGPT accurately describes the icon, and the explanation captures what we intended to convey through the use of that icon.

And while ChatGPT's output is accurate and meaningful, it is still too wordy to be useful as alternative text. Ideally, alternative text should be limited to 125 characters at most. Fortunately, we can ask ChatGPT to try again and limit the output for each icon description to 125 characters.

The maximum number of characters for alt text is 125 characters. Please revise accordingly. Thank you.

The alt text generated is now much more succinct and fits within our best practice guidelines for alt text. Let’s look at intrinsic motivation again: A thumbs-up sign representing encouragement and motivation.

The alt text is succinct and accurately captures what we want to convey. As a learning designer, it would have taken some time to draft the first version and additional time to revise the second version. And now, after seeing GenAI in action for this particular use case, I feel pretty confident it could help me develop my ability to write effective alt text.

Additionally, it could help accelerate and improve an important learning design process related to inclusive design principles and practices.

So, using ChatGPT to write alternative text to enhance digital accessibility is a promising use case to explore.

I encourage you to give it a try.