Long Description

Long descriptions should be formatted using headings, etc., just as with other accessible content.

• Make data tables available alongside charts and graphs. Tables provide a good alternative way of presenting the same information.

Long descriptions are too long to put in an alt text field, so there are two options for where it should go:

• Describe the image in the surrounding text. This benefits everyone reading the material.
• If space is limited, link to an appendix where you provide the long description. Just make sure that you also link back to the original location in the text for easier navigation within the document.
• Provide a companion document with each description clearly labeled with the original image title and page number from the original document.

Original images still need alternative text even when a long description is available, but it can be very basic. It can help to connect the image to the longer description somehow. For example:

• Alt Text: “Graph A, Page 3”
• Long Description: “Graph A on page 3 shows that…”

Context and situation becomes extremely important when writing descriptions for academic content, whether the image is being used in teaching content or assessing a student’s knowledge. Consider the following:

• How much background knowledge are students already expected to have on the subject?
• What are students expected to learn from the image?
• What are students expected to extract from the image?
• What information should not be given to the student so they can complete a required task?
• What information is necessary for the student to complete the task, whether learning or assessment?

Artificial Intelligence can be a powerful tool to help write good quality descriptions of complex academic content.

• Craft your prompt carefully for high quality descriptions.
• Review generated descriptions to ensure accuracy, appropriate knowledge levels and appropriate levels of detail for the task.

The following prompt templates were created using OpenAI ChatGPT 5.1, 2025.

Prompt Template:

Please generate a long, academically appropriate image description of the attached image. Follow these instructions exactly:

Goal:
Produce an objective, detailed, literal description suitable for academic documentation and screen reader accessibility.

Description Requirements:

• Identify the medium (photo, chart, illustration, painting, etc.).

• Provide a high-level overview, then progressively add details.

• Describe foreground, midground, and background elements.

• Identify all visible objects, people, text, or symbols.

• Include spatial relationships (“left of,” “behind,” “upper-right corner,” etc.).

• Describe scale, color, shapes, lighting, pattern, and texture.

• Indicate the count of objects where possible.

• Note partially visible or obscured elements.

Do NOT:

• Infer emotions, motivations, time period, culture, or context.

• Guess about unseen content.

• Add interpretation or subjective language.

Output Format:

1. A 3–5 sentence overview

2. A detailed bullet-point breakdown

Use For:

Any image containing math, symbols, formulas, reaction schemes, structures, or chemical equations.

Prompt Template:

Please convert all equations, formulas, and symbolic notations from the attached image into accessible MathML or ChemML.

Goal:
Produce a precise, faithful transcription of mathematical or chemical content for screen reader accessibility.

Instructions:

• Identify whether the content represents math or chemistry.

• Output MathML for math, ChemML for chemistry.

• Transcribe all content exactly:

  • Superscripts/subscripts

  • Fractions

  • Matrices, vectors, sets

  • Reaction arrows, stereochemistry, charges

  • Greek letters, operators, special symbols

• Maintain order, grouping, and layout.

• Include equation numbers/tags exactly as shown.

• For surrounding text that references formulas, transcribe it separately (not within MathML/ChemML).

• Mark ambiguous symbols as <unclear> without guessing.

Do NOT:

• Interpret or simplify formulas.

• Infer missing steps or context.

• Correct “errors” unless they are visually indicated.

Output Format:

1. Transcribed surrounding text (if present)

2. Each formula in a standalone MathML or ChemML code block

3. Maintain original sequence

Use for:

Graphs, charts, scientific figures, microscopy, CAD diagrams, engineering schematics, reaction schemes, lab setups.

Prompt Template:

Please generate a long, academically appropriate STEM-focused description of the attached image.

Goal:
Create a technical, objective description suitable for academic scientific documentation or research and screen reader accessibility.

Requirements:

• Identify the image type (graph, SEM micrograph, circuit diagram, chemical scheme, etc.).

• Describe axes, units, labels, scales, symbols, markers, legends.

• Transcribe visible numbers or text exactly.

• Describe spatial layout and relationships between components.

• For lab setups: list equipment, orientation, and arrangement.

• For chemical diagrams: describe bond types, functional groups, arrows, stereochemistry.

• For engineering or CAD: describe components, dimensions (if visible), and connections.

Do NOT:

• Infer scientific meaning, causes, conclusions, or problem-solving strategies.

• Fill in missing information or guess ambiguous elements.

Output Format:

1. Identification of STEM image type

2. High-level description

3. Technical bullet-point breakdown

4. List of all labels, text, symbols, and numbers

Please generate a detailed, academically appropriate description of the attached biological or veterinary image. Follow these instructions exactly:

Goal:
Produce an objective, technical, discipline-accurate description suitable for biomedical research, veterinary documentation, anatomical study, or archiving suitable for screen reader accessibility.

Requirements:

• Identify the type of biological image (anatomy photograph, radiograph, histology slide, tissue specimen, surgical field, microscopy image, behavioral observation, etc.).

• Provide an overall description first, then add finer anatomical or structural details.

• Describe visible tissues, organs, structures, or organisms using correct terminology (e.g., “proximal phalanx,” “epithelial layer,” “cranial,” “rostral,” “dorsal,” “medial,” etc.).

• Describe spatial relationships precisely (“the right kidney is lateral to the spine,” “the mass is located distal to the femoral head”).

• Transcribe all visible labels, tags, scale bars, magnification levels, or measurement markers.

• For microscopy: note coloration (true vs. stain artifact), cell types, visible morphology, layering, and patterns.

• For radiographs or ultrasound: describe views, orientations, densities, and visible structural boundaries without interpreting pathology.

• For animals: identify species only if explicitly visible; otherwise describe visible traits (fur color, limb shape, body features).

• For surgical or lab images: describe tools, positioning, incisions, and equipment.

• For ecological or behavioral images: describe visible actions objectively without inferring intent or emotion.

Do NOT:

• Diagnose, interpret pathology, or identify abnormalities beyond describing what is visually present.

• Infer species, sex, biological function, health status, or purpose unless unambiguously shown.

• Provide hypotheses or clinical judgments.

• Guess identities of tissues that are unclear; instead note ambiguity.

Output Format:

1. Identification of image type

2. A 3–6 sentence high-level overview

3. Detailed bullet-point breakdown of all structures, features, labels, or equipment

4. List of any text, labels, measurements, or scale markers exactly as shown

Use for: 

Standardized test diagrams, math problems, graphs, tables, multiple-choice materials, exam figures.
(Designed to avoid accidental hints.)

Prompt Template:

Generate a long, strictly neutral description of the attached exam image.

Goal:
Provide an accessible alternative format for an exam without giving hints or interpretations.

Requirements:

• Describe only the visible content.

• Transcribe text, labels, axes, numbers, answer choices exactly.

• Describe shapes, lines, angles, and spatial arrangement objectively.

• For graphs: specify axes, scales, plotted points, lines (without interpreting trends).

• For figures: identify geometric shapes, labels, and positions.

• For diagrams: list components and their relative placement.

Strictly Do NOT:

• Explain, analyze, or solve the problem.

• Provide definitions or background knowledge.

• Offer hints, simplifications, or relevant formulas.

• Interpret what the student should do.

Output Format:
A single, linear, objective description appropriate for a testing environment.