Source
Putting sign language AI into users’ hands
Google DeepMind introduced SL2T, a sign-language-to-text translation model now available in Gboard and Live Transcribe on Pixel 11. The feature initially supports translating American Sign Language (ASL) to English.
Why it matters
Deaf and hard of hearing users can now use sign language for searching and messaging instead of typing, increasing communication speed and digital accessibility.
The details
- The model was trained on over 100,000 hours across 50+ sign languages.
- Privacy is maintained by using pose landmark locations instead of raw video.
- Development included guidance from the AI Sign Language Advisory Committee (AISLAC).
Show entities and relationshipsHide entities and relationships
In this article
Topics
Companies
People
Organizations
Key connections
Google owns MediaPipe Holistic
Google develops the MediaPipe Holistic on-device tracking model
Sign Language-to-Text (SL2T) uses MediaPipe Holistic
SL2T uses MediaPipe Holistic on-device tracking to receive pose landmark locations
Users can sign into Gboard to ask Gemini to solve queries or execute tasks
Sam Sepah is a Deaf Googler at Google
Sam Sepah is related to Sign Language-to-Text (SL2T)
Sam Sepah conceptualized the sign language AI project
Google DeepMind is a partner of AI Sign Language Advisory Committee (AISLAC)
Google DeepMind established AISLAC to guide responsible deployment of sign language AI
Show 9 more connectionsShow fewer connections
Sign Language-to-Text (SL2T) is related to AI Sign Language Advisory Committee (AISLAC)
AISLAC co-authored a joint impact report for the release of SL2T 1.0
Sign Language-to-Text (SL2T) is built with Sign-to-Text
SL2T is an AI model built for sign-to-text translation
Sign Language-to-Text (SL2T) uses Computer Vision
SL2T relies on computer vision to interpret visual movements and body landmarks
Sign Language-to-Text (SL2T) is built with Machine Translation
SL2T performs machine translation from sign language landmarks directly into text
MediaPipe Holistic is built with Pose Estimation
MediaPipe Holistic tracks landmark points on the signer using pose estimation
MediaPipe Holistic uses Computer Vision
MediaPipe Holistic is an on-device computer vision tracking model
Pixel 11 integrates sign-to-text translation into the Gboard toolbar.
Pixel 11 uses Live Transcribe
Pixel 11 sign-to-text feature enables signers to express themselves when using Live Transcribe.
Live Transcribe uses Sign Language-to-Text (SL2T)
Live Transcribe supports sign-to-text translation powered by the SL2T model.
Related events
Google DeepMind Introduces Sign-Language-to-Text (SL2T) Translation Model
Get the weekly recap
The stories like this one, picked and explained — once a week, straight to your inbox.