IndesignPerfect

Accessibility is becoming an increasingly important part of professional publishing.

Whether you’re creating corporate reports, educational materials, ebooks, government publications, brochures, catalogs, or accessible PDFs, images often need meaningful alternative text so that people using screen readers can understand the information those images communicate.

Adobe InDesign can now help with this process using artificial intelligence.

With Generate Alt Text, InDesign can analyze an image and automatically suggest a description. Instead of writing every image description from scratch, designers can use AI-generated text as a starting point, review it, refine it, and incorporate it into the document’s accessibility workflow.

That last part is important.

AI can help generate alt text.

It cannot decide, entirely on its own, what an image means in the context of your publication.

Here’s how to use AI-generated alt text in Adobe InDesign effectively.

What Is Alt Text?

Alt text, or alternative text, is a written description associated with an image or other visual element.

People who can see the page receive information visually.

A screen-reader user may instead need that information communicated through text.

Imagine a corporate report containing a photograph of a new manufacturing facility.

A sighted reader immediately sees the image.

A useful alt-text description might be:

Exterior of the company’s new manufacturing facility with solar panels installed across the roof.

A screen reader can communicate that description to the reader.

Without meaningful alt text, important visual information may be lost.

Why Alt Text Matters in InDesign

InDesign is widely used to create documents that eventually become:

• accessible PDFs
• EPUB publications
• digital reports
• educational materials
• ebooks
• corporate publications
• government documents
• product documentation

Many of these formats are consumed with assistive technology.

Accessibility therefore shouldn’t be treated as something separate from design.

It is part of professional publishing.

A beautifully designed document that excludes some readers isn’t a completely successful document.

What Is Generate Alt Text in Adobe InDesign?

Generate Alt Text uses AI to analyze placed imagery and suggest an alternative-text description.

Instead of beginning with an empty description field, the designer can ask InDesign to generate a draft.

For example, suppose the placed image shows a cyclist riding through a forest.

AI might suggest something similar to:

A cyclist riding along a forest trail surrounded by tall trees.

The designer then evaluates that description.

Is it accurate?

Is it relevant?

Does it describe the information the image contributes to this particular publication?

If necessary, the designer edits it before using it.

The workflow is therefore:

Analyze → Generate → Review → Edit → Publish

Not:

Analyze → Generate → Publish automatically

That distinction is essential.

Why Use AI for Alt Text?

Writing one image description isn’t particularly difficult.

Writing descriptions for hundreds of images is another matter.

Consider a 250-page educational textbook containing:

• photographs
• diagrams
• illustrations
• maps
• screenshots
• charts
• infographics

Creating descriptions manually can become a substantial production task.

AI can reduce the amount of work involved in creating the initial draft.

Instead of starting from nothing, the designer or accessibility specialist starts with a suggested description.

Even when that description requires editing, it may still save considerable time.

How to Generate Alt Text in Adobe InDesign

The exact interface can evolve as Adobe updates InDesign, but the general workflow is straightforward.

Start by selecting the placed image or graphic that requires a description.

Access InDesign’s Object Export Options or the relevant alt-text controls available for the selected object.

From the Alt Text section, use the option to Generate Alt Text.

InDesign analyzes the image and produces a suggested description.

Read the generated text carefully.

If the description accurately represents the purpose of the image, you can use it as the basis for your final alt text.

If it is incomplete, misleading, overly detailed, or missing important context, edit it.

The final description should reflect the purpose of the image in your publication—not merely what the AI can recognize visually.

Example: A Simple Photograph

Imagine your document contains a photograph of Amsterdam’s canals.

AI might generate:

Canal in Amsterdam with historic buildings and bicycles along the waterfront.

For a travel brochure, that may be perfectly adequate.

But suppose the article specifically discusses Amsterdam’s seventeenth-century canal houses.

The more useful description might be:

Seventeenth-century canal houses lining an Amsterdam canal, with bicycles parked along the waterfront.

The difference is context.

AI sees an image.

The designer understands why that image appears in the publication.

Context Is More Important Than Description

This is one of the most important principles of good alt text.

The goal isn’t necessarily to describe every visible detail.

The goal is to communicate the meaning or function of the image.

Consider a photograph showing a wind turbine.

In an article about renewable energy, appropriate alt text might be:

Wind turbine generating renewable electricity in a rural landscape.

But if the same photograph appears in an article discussing turbine blade design, the relevant description may be:

Three-bladed horizontal-axis wind turbine showing the long aerodynamic blade design.

Same image.

Different context.

Different alt text.

AI cannot always infer that editorial purpose from the pixels alone.

AI Alt Text Is a Starting Point

Think of Generate Alt Text as an assistant creating the first draft.

That draft may be:

• completely usable
• mostly correct
• too generic
• too verbose
• missing context
• factually uncertain
• inappropriate for the image’s purpose

Your job is to decide which.

This is why human review remains essential.

The AI saves time by eliminating the blank field.

The human provides editorial judgment.

Don’t Make Alt Text Unnecessarily Long

Good alt text is generally concise.

A screen-reader user shouldn’t have to listen to a 150-word description of a decorative office photograph unless all those details genuinely matter.

Consider:

Photograph showing a modern office containing six desks, eight chairs, three computer monitors, four plants, two large windows, white walls, wooden flooring, hanging lights, shelves, notebooks and several other office objects.

Technically detailed.

Not necessarily useful.

A better description might simply be:

Modern open-plan office with desks, large windows and indoor plants.

The right level of detail depends on the image’s purpose.

Avoid Starting Every Description with “Image of”

Screen readers already identify images as images.

That means descriptions such as:

Image of a woman using a laptop.

often contain redundant information.

Usually:

Woman working on a laptop at a desk.

is more efficient.

There are exceptions where identifying the type of visual matters.

For example:

Diagram showing the stages of the publishing workflow.

or:

Screenshot of the InDesign Paragraph Styles panel.

The content type is relevant because it helps the reader understand what is being communicated.

Decorative Images May Not Need Alt Text

Not every image needs a meaningful description.

Some images exist purely for decoration.

Examples might include:

• ornamental flourishes
• background textures
• decorative lines
• visual separators
• non-informative shapes

If an image communicates no meaningful information, forcing a screen-reader user to listen to a description may actually make the document less accessible.

Decorative content should generally be identified appropriately so assistive technology can ignore it.

The important question is:

Would a reader lose meaningful information if this image disappeared?

If the answer is no, the image may be decorative.

Functional Images Need Functional Descriptions

Some graphics perform an action rather than simply communicate visual information.

For example, an icon might function as a button or navigation control.

In those cases, alt text should describe the function, not merely the appearance.

Poor:

Blue square containing a white arrow.

Better:

Next page.

The second description tells the user what the control does.

This principle is particularly important in interactive documents and digital publishing.

Charts Need Special Attention

AI-generated alt text becomes more challenging with charts and graphs.

Suppose an annual report contains a bar chart showing revenue growth from 2022 through 2026.

A generic AI description might say:

Bar chart showing company revenue over several years.

That identifies the visual.

But it may not communicate the important information.

A more useful description might be:

Bar chart showing annual revenue increasing from €12 million in 2022 to €21 million in 2026, with growth recorded each year.

Now the reader receives the key takeaway.

Complex charts may require even more detailed treatment elsewhere in the document.

AI can help identify the image, but accessibility often requires understanding the data.

Diagrams Also Require Human Judgment

Technical diagrams create similar challenges.

Imagine a diagram showing a cloud architecture.

AI might recognize:

Diagram with boxes connected by arrows.

That isn’t very useful.

A subject-matter expert might describe it as:

Architecture diagram showing users connecting through a web application to an API layer, which communicates with the application database and cloud storage.

Again, the important information isn’t merely what the image looks like.

It’s what the diagram means.

Screenshots Need Purpose-Oriented Alt Text

Software documentation frequently contains screenshots.

Suppose an InDesign tutorial includes a screenshot of the Paragraph Styles panel.

Weak description:

Screenshot showing an Adobe InDesign window.

Better:

InDesign Paragraph Styles panel showing Body Text, Heading 1, Heading 2 and Caption styles.

The second description communicates why the screenshot exists.

This is particularly relevant for websites and publications teaching software.

AI Can Help with Large Image Collections

Where Generate Alt Text becomes particularly valuable is scale.

Imagine producing a product catalog containing 800 photographs.

Creating 800 descriptions manually is a major task.

AI can generate first drafts.

The production team can then review them systematically.

A workflow might look like:

AI generation → Editorial review → Accessibility review → Final export

This can be significantly faster than writing every description from scratch.

InDesign Can Flag Potentially Outdated Alt Text

One particularly useful development in recent InDesign versions is the ability to indicate when alt text may need updating after an image changes.

Suppose you place a photograph showing three people at a conference.

You generate alt text:

Three speakers discussing artificial intelligence during a conference panel.

Later, you crop the photograph so only one speaker remains visible.

The original description is now inaccurate.

Recent InDesign updates can provide a visual indication that existing alt text may have become outdated after the image is cropped or repositioned.

You can then regenerate or manually update the description.

This is a subtle but important accessibility improvement.

Why Image Changes Matter

Designers modify image crops constantly.

It is part of normal layout work.

You might:

• zoom into an image
• reposition the subject
• crop out background elements
• replace an image
• change the visible area

Historically, alt text could easily become disconnected from the final visual.

The designer changes the crop but forgets the accessibility metadata.

By linking image changes more closely to alt-text review, InDesign makes accessibility part of the normal editing workflow.

Accessibility Shouldn’t Be a Final-Step Checklist

This represents a larger shift.

A common production model has been:

Design everything → Finish document → Make it accessible

That approach can create problems.

Accessibility issues discovered at the end may require substantial corrections.

A better workflow is:

Design + Accessibility together

As images are placed, consider alt text.

As document structure is created, consider reading order.

As styles are built, consider semantic structure.

As colors are chosen, consider contrast.

As digital content is exported, consider how it responds to text resizing.

Accessibility becomes part of the document architecture.

Generate Alt Text and EPUB

Alt text is particularly important when exporting reflowable EPUB content.

Ebook readers may use screen readers or other assistive technologies to communicate visual information.

Meaningful image descriptions therefore become part of the reading experience.

InDesign’s accessibility and EPUB improvements increasingly work together.

Designers creating ebooks should test their exports rather than assuming that adding alt text automatically makes the publication accessible.

A complete EPUB accessibility workflow involves more than image descriptions.

Generate Alt Text and Accessible PDF

The same principle applies to PDF.

Adding descriptions to images is an important part of accessible PDF production, but it is only one part.

Accessible PDF documents may also require attention to:

• document tags
• heading structure
• reading order
• table structure
• language
• bookmarks
• form fields
• color contrast
• link descriptions
• document metadata

Generate Alt Text helps solve one specific accessibility task.

It doesn’t automatically make an entire PDF accessible.

Review AI Descriptions for Bias and Assumptions

AI-generated descriptions can sometimes make assumptions that aren’t necessary or appropriate.

For example, an AI system might infer:

• someone’s occupation
• someone’s relationship to another person
• an emotional state
• a location
• an activity

when those things aren’t actually clear.

Alt text should avoid unsupported assumptions.

Instead of:

Business executive celebrating a successful sales meeting.

the visually supportable description may simply be:

Person standing beside a presentation screen while colleagues applaud.

Describe what matters and what can reasonably be established.

Be Careful Identifying People

If an image contains a known person, AI may or may not correctly identify them.

Don’t assume the generated identification is accurate.

If the person’s identity matters, verify it.

If it doesn’t matter, a generic description may be more appropriate.

For example:

Speaker addressing an audience from a conference stage.

may be perfectly adequate.

Again, context determines what the reader needs.

Don’t Include Information That Sighted Readers Don’t Receive

Alt text shouldn’t normally provide hidden information unavailable from the visual.

Suppose a photograph shows a new office building.

Don’t write:

Company’s €25 million headquarters completed in March 2026.

unless that information is actually communicated by the image or surrounding content and is appropriate to the context.

Alt text should provide equivalent access—not an entirely different information experience.

Avoid Repeating Nearby Text

Sometimes an image’s meaning is already explained immediately beside it.

For example, a chart may have a detailed caption describing the trend.

Repeating exactly the same information in alt text may create unnecessary duplication for screen-reader users.

Look at the image in context.

Ask:

What information does the reader still need from this image?

Accessibility is about the complete reading experience, not isolated metadata fields.

A Practical AI Alt-Text Workflow

For professional InDesign production, consider this workflow:

Step 1: Place the image

Build the layout normally.

Step 2: Determine whether the image is meaningful

If it is purely decorative, treat it appropriately.

Step 3: Generate alt text

Use InDesign’s AI capability to create a first draft.

Step 4: Read the description

Don’t approve it automatically.

Step 5: Compare it with the actual image

Check whether the visible content is described accurately.

Step 6: Consider the editorial context

Ask why the image appears on this page.

Step 7: Remove unnecessary detail

Keep the description focused.

Step 8: Add missing meaningful information

Make the description useful to the reader.

Step 9: Review after cropping or repositioning

Pay attention to InDesign’s indication that alt text may need updating.

Step 10: Test the exported publication

Accessibility doesn’t end inside InDesign.

This approach combines AI efficiency with human responsibility.

A Simple Alt-Text Checklist

Before approving generated alt text, ask:

Is it accurate?

Does it correctly describe the visible image?

Is it relevant?

Does it communicate why the image matters here?

Is it concise?

Can unnecessary words be removed?

Does it avoid assumptions?

Is everything stated reasonably supported?

Does it duplicate nearby text?

If so, can the description be simplified?

Has the image changed?

If it was cropped or replaced, does the description still apply?

Is the image actually decorative?

If yes, meaningful alt text may be unnecessary.

These questions can catch many common problems.

AI Doesn’t Replace Accessibility Expertise

Generate Alt Text can dramatically accelerate one part of the accessibility process.

But accessibility is broader than AI image descriptions.

Professional accessible publishing requires understanding:

• semantic structure
• document navigation
• reading order
• contrast
• typography
• tables
• links
• forms
• responsive text
• export formats
• assistive technology

AI is useful because it reduces repetitive work.

It doesn’t remove the need to understand the standards and the reader’s experience.

Why This Feature Matters

Generate Alt Text isn’t as visually spectacular as Text to Image or Generative Expand.

It won’t produce dramatic images for social media demonstrations.

But from a professional publishing perspective, it may be one of InDesign’s most meaningful AI applications.

Why?

Because it addresses a real production bottleneck.

Creating useful descriptions for large numbers of images takes time.

AI can accelerate the first-draft stage.

That gives designers and accessibility specialists more time to concentrate on whether those descriptions are actually useful.

The Bigger Picture: AI as an Accessibility Assistant

Generate Alt Text also demonstrates a particularly valuable role for AI.

Instead of replacing creative decisions, AI can assist with tasks that are necessary but repetitive.

The ideal relationship looks like this:

AI handles initial analysis.

Human provides context.

AI accelerates production.

Human verifies quality.

This pattern is likely to appear increasingly throughout professional publishing.

Final Thoughts

AI-generated alt text can make accessible publishing in Adobe InDesign faster and more manageable.

But the feature works best when its purpose is understood correctly.

It isn’t:

“Let AI handle accessibility.”

It is:

“Let AI help create the first draft so I can concentrate on making the description useful.”

That’s an important distinction.

A computer can recognize that an image contains mountains, a lake and a person.

A designer understands that the image appears in a travel article about hiking in the Canadian Rockies.

A computer can recognize that an image contains bars and numbers.

An editor understands that the chart’s important message is a 40% increase in revenue.

A computer can generate words.

A human determines which words matter.

That combination is where Generate Alt Text becomes genuinely useful.

As Adobe continues integrating AI into InDesign, accessibility is becoming increasingly connected to the everyday design workflow.

And that may ultimately be far more important than simply making alt text faster to write.