Name the category, not every detail

Use specific high-level labels such as sexually explicit adult imagery, graphic injury, flashing light, or sudden loud audio. Vague phrases such as 'mature content' may not give a reader enough information, while lurid descriptions can reproduce the harm the warning is meant to prevent.

Separate content from conduct. A page documenting a historic explicit artifact can say what the artifact contained without presenting it. The warning should match what the next action will actually load.

A useful warning preserves reader choice

  1. A warning belongs before the browser requests or previews sensitive media.

  2. Calm category labels inform better than vague or sensational language.

  3. Keyboard access, readable text, visible focus, and predictable navigation are part of the warning.

Place the choice before the request

Do not preload the full image behind a blur, fetch it for a hover preview, or include it in social metadata if the visitor has not opted in. Network requests can expose the file to browser history, caches, filters, link unfurlers, and assistive tools even when CSS hides it visually.

Use an ordinary link or button with a clear destination and no automatic navigation. If activation opens a new tab or external site, disclose that behavior so keyboard and screen-reader users are not unexpectedly moved into another context.

Make the warning accessible

Provide real text rather than text baked into an image. Keep the control keyboard-operable, give it a visible focus state, and ensure its accessible name describes the action. Color and an icon can reinforce the warning but should not be the only signal.

Avoid countdowns, motion, shame-based language, or a large primary button that pressures visitors toward exposure. Declining should be as easy as proceeding, and the back path should remain obvious.

Choose documentation over redisplay

Researchers often need identity, dates, filenames, page relationships, and reception - Not the media itself. A non-graphic description and source trail can satisfy those needs while keeping the archive broadly usable.

If a publisher does host sensitive primary material for a legitimate research purpose, isolate it from general navigation, prevent automatic previews, label age or legal restrictions where applicable, and document the editorial reason for inclusion.

Warning placement and accessibility questions

Is blurring an image enough warning?

Not if the file loads automatically or the blur still reveals meaningful detail. A safer design places a text choice before any request for the sensitive media.

How specific should a content warning be?

Name the relevant category clearly enough for an informed choice, while avoiding sensational detail that itself becomes distressing or promotional.

Should a warning open sensitive content in a new tab?

Only with advance notice and a clear reason. Unexpected new contexts can disorient users and weaken the browser's normal back path.

Sources and further reading

  1. W3C guidance on warning before new windows
  2. WCAG 2.1 predictable-behavior guidance
  3. W3C understanding link purpose in context

Sources are provided for verification. External destinations may contain material GOAT.CX does not host or control. Review the source and corrections policy and the maintained GOAT evidence ledger.

Publication record

First published . Last substantive review . The updated date changes only after a source, factual, or explanatory revision - Not an automated timestamp refresh.