Accessibility is no longer optional - but at scale, it’s also not simple.

Yet one of the most common accessibility failures across Drupal sites remains surprisingly simple: image alt text. Missing, vague, duplicated, or technically present but semantically meaningless alt attributes quietly break WCAG compliance, harm SEO, and exclude users who rely on assistive technologies.

Large Drupal sites often contain thousands of media items created over the years by different editors, migrations, and integrations. For these sites, manually auditing image accessibility is slow, inconsistent, and prone to error. And while AI promises quick answers, blindly auto-generating alt text introduces a different risk - loss of editorial control.

At Specbee, we believe accessibility and AI should work together, but not at the cost of trust.

That belief led us to build Media Accessibility Audit, a Drupal 11 module designed to audit, score, guide, and improve image accessibility - without ever mutating content automatically.

The real challenge with image accessibility

Most Drupal sites struggle with image accessibility for a few key reasons:

• Alt text is often added late - or not at all
• Editors use placeholders like “image” or “photo”
• Media libraries grow faster than audits can keep up
• Accessibility reviews happen too late in the release cycle

Alt text might be missing, vague, duplicated, or technically present but semantically meaningless.

Traditional solutions fall into two extremes:

• Manual audits that don’t scale and only tell you something is wrong
• AI auto-generation that overwrites content without context or oversight

Neither is ideal.

Accessibility needs human judgment, but humans need better tools.

A different philosophy: AI as guidance, not automation

AI Media Accessibility Audit was built to solve exactly this problem - without taking control away from editors or administrators.

Instead of auto-writing content, the module focuses on auditing, scoring, and assisting decisions, using AI responsibly and transparently.

Its philosophy is simple:

• Accessibility improvements should be auditable, reviewable, and intentional
• AI should assist editors - not replace them

The module emphasizes:

• Analysis, not silent changes
• Transparency, not black-box behavior
• Editorial choice, not forced automation

AI-generated alt text is always:

• Presented as a preview
• Clearly labeled with confidence indicators
• Applied only when explicitly confirmed

Nothing is saved automatically. Ever.

Why AI Media Accessibility Audit?

Most accessibility tools fall into one of two extremes:

❌ They only tell you something is wrong, without helping you fix it
❌ Or they automatically overwrite content, removing human oversight

Media Accessibility Audit takes a different approach:

✔ Audits existing alt text quality
✔ Detects meaningful vs meaningless alt text
✔ Scores accessibility per media item
✔ Provides read-only AI suggestions
✔ Applies changes only with explicit confirmation

This makes it suitable for editorial teams, accessibility reviews, and enterprise governance workflows.

What the AI Media Accessibility Audit module does

Once enabled, the module introduces a dedicated Media Accessibility Audit dashboard inside Drupal.

Media Accessibility Audit Dashboard (UI)

Path:
/admin/content/media-accessibility-audit

This dashboard is the central hub for reviewing and improving image accessibility across your site.

The dashboard displays media previews, accessibility scores, detected issues, and AI-generated alt text suggestions with confidence indicators.

From the dashboard, administrators can:

• Review media items with accessibility scores
• See detected alt-text issues (missing, vague, generic, or low-quality)
• Preview AI-generated alt text suggestions (read-only)
• Apply suggested alt text to individual media items
• Refresh AI suggestions per media item

🛡️ Important:
No content is modified unless an administrator explicitly confirms the change.

Accessibility scoring that reflects reality

Each media item is analyzed based on its stored alt text, not AI output.

Each media item receives:

• A Pass / Fail status
• A numeric accessibility score
• A clear list of detected issues

For example:

• Missing alt text
• Alt text present but meaningless
• Alt text mismatched with media context

This ensures scoring is accurate and trustworthy and makes it easy to prioritize fixes instead of treating all issues as equal.

Responsible AI suggestions (not auto-writing)

When an AI provider is configured, the module can generate AI-powered alt text suggestions that are:

• WCAG-conscious
• Context-aware
• Length-controlled
• Clearly marked as suggestions

If the configured AI model supports vision input, the image itself is used. Otherwise, suggestions are generated using media and content context - with confidence adjusted accordingly.

Each suggestion includes:

• A confidence indicator (low/medium/high)
• A clear signal when vision models were used
• Provider badges

Key design decisions:

• Suggestions are read-only previews
• Editors can review, edit, or ignore them
• Suggestions are never applied automatically

This ensures AI assists - it does not replace - human judgment.

Single media review

Each media item can be reviewed independently for precise editorial control.

Editors can:

• View detected alt-text issues
• Preview AI-generated suggestions
• See confidence indicators and vision usage
• Apply changes manually if desired

This supports editorial workflows where accuracy and context matter more than speed.

Bulk apply alt text (UI)

For larger cleanup tasks, Media Accessibility Audit supports bulk operations directly from the dashboard.

Administrators can:

• Select multiple media items
• Review AI-generated alt text suggestions
• Check confidence indicators before applying changes
• Apply suggestions in bulk with explicit confirmation

⚠️ Bulk actions never bypass editorial consent.

Batch audit for large media libraries

For sites with large media libraries, the module uses Drupal’s Batch API to process audits safely.

From the dashboard, you can:

• Start a batch audit
• Process media items in safe chunks
• Avoid timeouts on large datasets
• See results appear progressively

This ensures performance stability even on enterprise-scale sites.

Drush command (CLI)

Media Accessibility Audit also integrates with Drush, making it suitable for automation and advanced workflows.

Run a standard audit:

drush media_accessibility_audit:audit

Batch-safe mode for large sites:

drush media_accessibility_audit:audit --batch

CLI audit output highlights missing or low-quality alt text, making it suitable for:

• CI pipelines
• Cron-based accessibility reviews
• Pre-launch accessibility checks
• Enterprise-scale media libraries

Built for trust, not automation

Media Accessibility Audit is intentionally not an auto-fix tool.

That’s why:

• Stored alt text is always the source of truth
• AI suggestions never silently modify content
• Editors stay in control at every step

By supporting UI-based reviews, bulk actions, batch processing, and CLI automation, Media Accessibility Audit fits naturally into your Drupal projects - from small editorial teams to enterprise platforms.

Final thoughts

Media Accessibility Audit helps teams move from guessing to knowing:

• Which media items have problems
• Why do those problems exist
• How to fix them safely

By keeping AI suggestions transparent, optional, and context-aware, Media Accessibility Audit helps teams improve WCAG compliance, maintain editorial trust, scale accessibility efforts responsibly, and avoid risky auto-generated content.

Accessibility deserves more than automation - it deserves care.