Complex Images Descriptions

Use the alt attribute for a brief label (e.g., “Bar chart showing enrollment by department, see text description below”).

Provide a long description - a more detailed description of the image.  Often necessary for complex visualizations such as graphs, bar charts, etc.

• Add the full explanation in surrounding text, a figure caption, or a linked description page.
   
• the long description can go under the image, link out to a different page, or in speaker notes for slides

Summarize the overall message first, then break down details (e.g., trends in a graph or steps in a diagram).

Don’t repeat the same text in multiple places as the description is already long enough, we don't want the student to get confused.

• Efficiency: Quickly create draft descriptions instead of writing from scratch
• Idea Generation: Helps you think about what information an image conveys and what students may need.
• Support for complex content: AI tools can describe graphs, charts, diagrams, or even mathematical notation.

These AI tools can produce draft descriptions when provided with an image and a detailed prompt. Some tools that you can explore:

1. Blackboard Ally:

   Rutgers faculty also have access to Blackboard Ally, an accessibility tool integrated into our LMS. Ally now includes a feature that can generate AI-based alternative text suggestions for images uploaded to course content. 

   Learn More:

   1. Add Image Descriptions with Ally (Blackboard Help)
   2. Blackboard Ally Release Notes – July 17 (includes updates for handwriting recognition)

2. Microsoft Copilot:

   Copilot (available in Microsoft 365 apps) can suggest descriptions for images uploaded into Word, Powerpoint, or Outlook. It's a convenient option for faculty already working in Microsoft environments.

3. Google NotebookLM:

   NotebookLM is designed by Google to summarize and explain large amounts of text. Faculty can upload lecture notes, slides, or readings, and then ask it to produce image descriptions consistent with the course context. This ensures that alt text connects directly to what students are learning.

4. Image Accessibility Creator (from ASU):

   Developed by Arizona State University, this tool allows users to create and refine image descriptions with accessibility in mind. It is particularly useful for faculty who want a structured approach to drafting alt text.

Other GenAI tools, such as ChatGPTand Google Gemini, can also generate draft alt text. These work best when faculty provide detailed prompts and context from course materials. As always, descriptions should be reviewed and refined to ensure they are accurate and meaningful for students.

• Write Detailed Prompts: AI works best when given clear instructions. Instead of asking “Write alt text,” try a prompt like:

“Act as an accessibility expert and provide alt text for the following image. Follow WCAG standards and make sure the description is meaningful to a blind or low vision student.”

• Provide Context: Context plays a key role in accessibility. You can upload or copy in a larger portion of your study material (slides, textbook excerpts, lab manuals, etc.) so the AI understands how the image fits into the lesson. A diagram used in a biology lecture should be described differently than the same diagram in a history of science course.
• Always review and edit. Even with strong prompts and context, AI may still mislabel or over-describe elements. Read the draft critically and make adjustments so the description is accurate and relevant to your course.
• Focus on purpose, not appearance. Focus on what students need to learn from the image, not every visual detail. For example, instead of “a red line going upward from left to right,” say “graph showing steady growth in enrollment from 2015 to 2025.”