Ally Now Supports AI-Generated Descriptions for Images
Leverage AI to provide inclusive resources for students
Recent updates to federal guidelines for digital accessibility underscore the critical need for educators to ensure that all course materials are accessible to every student.
One of the most prevalent accessibility issues in digital course materials is the lack of proper descriptions for images. According to Anthology, nearly one-third of all images flagged by Ally are simply marked as decorative without any descriptive text, which creates a barrier for students who rely on descriptions for their studies.
Properly tagging images not only meets federal guidelines, but also plays a crucial role in creating an inclusive learning environment. Alternative descriptions serve as vital information for students who use screen readers, providing them with textual explanations for images that otherwise visually conveyed. This practice not only supports students with visual impairments, but also benefits those with cognitive disabilities who might struggle to understand the context of images with additional context.
Ally will flag images that lack a description in a Blackboard course. Instructors can enter a description themselves, mark the image as decorative, or leverage AI tools, powered by Microsoft Azure AI Vision, to scan the image and generate a suggested description. Instructors retain full control over the image and can edit the content to make corrections or clarify the information before saving it.
Image 1: Example of Ally's AI generated image description.
Note: Alternative descriptions are not applied automatically. Each suggestion requires the instructor’s review and approval to ensure that the final output aligns with the instructor’s original purpose for that image.
If the alt text does not meet the instructor’s needs, such as a complex image or diagram, ChatGPT and Microsoft Copilot provide excellent tools to create detailed alternative text and long descriptions.
For example, this prompt was shared by the International Ally Users Group. It can be customized for the uploaded image according to context, function and prospective audience (e.g., students).
“You are a university-level instructor working to ensure images in your online course materials conform to Web Content Accessibility Guidelines at Level Double A. You need to accommodate for disabilities and impairments such as blindness, low vision and color vision deficiency to ensure equitable access to module content for all students.
"You will be provided with an image. Please return a succinct title, alternative text (alt text), and a long description. You will also separately list verbatim all words that appear in the image.
"The alt text should be no more than two sentences and should briefly describe the meaning conveyed by the image.
"The detailed description may be one or more paragraphs and must describe the image in great detail so that anyone who does not see it can gain a full understanding of everything in the image. Please format the detailed description such that it will be easily scannable. Use headings, paragraph breaks, and lists as appropriate. Use active voice.”
Prompt created by Purdue University and adapted by Matthew Deeprose (Southampton University)
When we integrate comprehensive and accurate alt text for all images, we adhere to accessibility standards and legal requirements. We also foster an inclusive and supporting learning environment where every student has the opportunity to succeed.
Posted: August 13, 2024, 9:26 AM
