Google DeepMind launches SL2T sign language translation for Pixel 11

Google DeepMind has introduced SL2T (Sign-Language-to-Text), an AI model designed to translate American Sign Language (ASL) into English text in real time. The technology is now available on Pixel 11 devices through the Gboard keyboard app and the Live Transcribe service.

Users can utilize sign language to perform tasks such as web searches, drafting messages, writing documents, or interacting with the Gemini AI assistant, rather than typing manually. The model was trained on over 100,000 hours of data spanning more than 50 sign languages, with ASL accounting for approximately one-quarter of the training capacity.

To prioritize user privacy, the system does not store raw video footage. Instead, an on-device model called MediaPipe Holistic tracks 130 coordinate points across the hands, face, and torso, converting movements into geometric data before discarding the video. While the model shows improved accuracy over previous versions, Google noted challenges with rare signs, rapid fingerspelling, and certain complex grammatical structures.