Twitter says its ALT badge for images and improved image descriptions are now available globally, after testing the features in March
Kim Lyons / The Verge :
Context & Ripple Effects
Twitter has been building toward this for six years: it first let users attach descriptive text to images in its mobile apps back in 2016, but that metadata stayed invisible to most readers. The ALT badge tested in March changes that by surfacing a visible marker on any image that carries a description, with the text revealed on tap.
The global rollout closes the loop on that test and puts Twitter level with Instagram, which has offered AI-generated or user-written photo descriptions for visually impaired users since 2018.
First-order effects
- Blind and low-vision users can now see at a glance which images in their timeline carry descriptions instead of tapping blindly, and sighted users get a one-tap way to read them.
- Authors who write alt text finally get visible confirmation their effort is discoverable — the badge turns an invisible field into public metadata.
Second-order effects
- Visibility creates accountability: with badges exposing which posts lack descriptions, Twitter followed up months later by testing upload-time reminders nudging users to add them — pressure that only makes sense once absence is publicly legible.
- Instagram's head start with automated descriptions raises the bar from manual-only alt text toward AI-assisted captioning as the expected baseline across major platforms.
Third-order effects
- If the pattern holds, accessibility metadata migrates from an opt-in authoring chore to a surfaced, platform-enforced layer of every image post — the same direction of travel as efforts to attach verifiable context to media at the capture layer.
- Platforms that surface image context gain a moderation and trust dividend too: described images are auditable images, which matters as regulators scrutinize how platforms handle visual misinformation.
The trend: Social platforms are moving image accessibility from hidden optional fields to visible, increasingly automated layers of the post itself.