Google’s AI-Powered Accessibility Push: 7 Key Upgrades You Need to Know

From Expressive Captions to African Language Support

On Global Accessibility Awareness Day (GAAD), Google announced updates for Android and Chrome that use AI to assist users with vision, hearing, or speech needs. The upgrades include Gemini-powered screen descriptions and open-source speech tools, reflecting a focus on personalization and global inclusivity. Here’s what’s new.

“Accessibility isn’t a checkbox. It’s about designing for the edges, where needs are most nuanced,” says a Google engineer involved in Project Euphonia.

First, TalkBack’s Gemini integration improves image descriptions. Users can now ask contextual questions about screen content, expanding on last year’s auto-generated alt text. Android 15+ introduces Expressive Captions, which stretch words like “yesssss” to match speech duration and label ambient sounds (e.g., [door creaking]). Currently limited to English in four regions, this feature enhances real-time transcription.

The Open-Source and Global Play

Google.org partnered with University College London to launch the Centre for Digital Language Inclusion (CDLI), focusing on speech recognition for 10 African languages—addressing the dominance of Eurocentric models. Project Euphonia’s GitHub release allows developers to create custom audio tools for non-standard speech patterns, such as ALS or accented dialects. “Democratizing these models accelerates solutions we’d never ideate in a lab,” notes the Euphonia team.

Chromebooks also receive updates: Face Control (head tracking for navigation) and Reading Mode (distraction-free text) are now available, alongside full accessibility tools in Bluebook, the digital SAT platform. Chrome’s PDF OCR enables screen readers to process scanned documents, while Android’s Page Zoom adjusts text size without disrupting layouts—fixing a long-standing issue.

“The next frontier? AI that adapts to individual disabilities, not the other way around,” predicts a CDLI researcher.

With 1.3 billion people globally living with disabilities, these updates—though incremental—point toward a future where assistive tech is proactive rather than reactive. The question is whether competitors will adopt similar open-source approaches or let Google’s frameworks set the standard.