DeepMind’s new sign-language-to-text (SL2T) model debuts with the launch of the Pixel 11 family, embedded inside Gboard and Live Transcribe to let deaf and hard of hearing users sign to their phone in places where they would normally type. The model supports ASL to English transcription at launch, trained on more than 100,000 hours of multilingual sign language data, with approximately a quarter of the dataset representing American Sign Language. SL2T does not interpret raw video; a separate on-device model converts footage into a wireframe of geometric coordinates sent to Google’s servers, designed to protect user privacy. It translates wireframes directly into text, avoiding intermediate glosses, which DeepMind says fail to capture non-linear aspects of sign languages. The company notes global diversity, with more than 70 million deaf and hard of hearing people communicating in some 200 sign languages, and plans to support more languages and devices in the future.