dmkeeler6

1. What surprised you?

While I have been excited for the opportunity to learn new methods and tools through this course, making technology-based instructional materials more accessible was not a piece I had thought much about. Not because I am not interested in making instructional materials more accessible, but simply because, in education, we often think about the use of technology itself as being what makes the learning accessible. Of course, this is a fallacy and just like any other teaching method, it is important to be intentional about accessibility when creating any instructional materials. While some web content accessibility guidelines, such as color choice and text size were choices I have thought about in the past, others were new to me. For example, the importance of not only creating headings and labels (WCAG Success Criterion 2.4.6), but also considering whether those headings are navigable and machine-readable was not something I had previously considered. The Minga Installation and Onboarding Guide I created was the first time I had used the headings feature in Microsoft Word. This gave me a new perspective on the importance of using pre-made features in Microsoft Word and that these may be useful beyond convenience. Another feature I had not considered before was creating alt-text for images (WCAG Success Criterion 1.1.1). Before the accessibility lesson, I did not realize that I could create alt-text in Microsoft Word for images and I hadn’t considered the importance of creating alt-text that not only describes an image but describes what the user should do. For example in the Minga Installation and Onboarding Guide artifact I created, it was important to explain what steps were being illustrated by a screenshot instead of focusing on the colors or design of the image. I will be continuing to create alt-text for my images moving forward and being mindful of whether a user who cannot access the image itself would be able to accomplish the same goal as a user who can access the image.

2. What patterns did you notice in AI-generated alt text?

I utilized AI-generated alt text as a resource in my Minga Installation and Onboarding Guide artifact and my AI Alt-text Critique blogpost. In the onboarding guide I used the AI generator in Microsoft Word and for the critique I used Google Gemini. Despite using different AI generators for each artifact, I saw very similar patterns to the AI-generated text. The AI-generators were excellent at describing what an image looked like and explaining a page layout in words that I would not have come up with on my own. For example, when asked to generate alt-text for an image of a GoGuardian dashboard with a variety of pages a user can navigate, Google Gemini described it as a “left-hand navigation menu” which was a descriptive, yet simple to follow, explanation of the menu. When using Microsoft Office’s AI generated alt text, it did a great job describing how the Apple Safari web browser icon looks: “Icon featuring a circular compass design with a blue gradient background and white tick marks around the edge.” However, neither AI software was able to effectively explain the purpose of the image within the document. For example, in the GoGuardian dashboard screenshot, the reason the “Classroom” button was circled in the left-hand navigation menu was to cue the user to click on the button to get to the Classroom dashboard page, but the AI-generated text did not provide that information. 

3. When should AI be used for alt text?

The way I view AI use for alt-text is similar to the way I view AI use in general. AI can be a very useful resource when creating web-based materials. It can create prose efficiently and stream-line text organization and descriptive based writing. However, I do not think AI created material should ever be taken as flawless and therefore should always be reviewed by a human before being finalized. The patterns I describe in the question above illustrate the need for caution and review when using AI for alt text. If I had used the AI generated alt-text for my images without revisions, the alt-text for my artifacts would not have provided the intended information for the user. For example, the alt-text for my screenshot of the “Scenes” dashboard in GoGuardian in my AI alt-text critique blogpost, described the page as having two cards that offer options to configure website access with “Allowed Website List” and “Blocked Website List”. However, it missed an explanation of the features found in those options. Without this information, the user cannot make an informed decision about which scene they are needing to create. Without revision, the alt-text for the image would not have created better access for the user. AI generated alt-text is very useful when an image is included just for the picture it provides and a description of the image is needed. However, it should not be used when an image needs to prompt a user to complete a task. 

4. How will this change your future design work?

The accessibility lesson was very eye-opening for me. I would like to challenge myself to consider the Web Content Accessibility Guidelines (WCAG) when creating artifacts. This lesson definitely showed me that I should not assume an artifact is accessible and instead need to use resources such as Microsoft Word’s accessibility check, Web Aim contrast checker, or free web accessibility evaluation tools, such as Wave to ensure I am creating tools that can be accessed by everyone. I will do my best to remove unnecessary barriers so that all learners can succeed and grow.