Google DeepMind has shipped its sign-language-to-text AI model, called SL2T, inside two Android apps on the Pixel 11, giving Deaf and hard-of-hearing users the ability to sign in American Sign Language and receive an English text translation in real time. It is the first time sign language AI has reached a shipping phone feature.
What the feature does and where it lives
SL2T powers sign-to-text dictation in Gboard and Live Transcribe on Pixel 11, starting with American Sign Language to English.
Through Gboard, users grant the keyboard permission to use the phone’s camera to translate ASL into English text in real time. The feature works inside any app that accepts text input, covering web searches, drafting emails and text messages, editing documents, and prompting Google’s Gemini assistant.
In Live Transcribe, users can sign their side of a conversation rather than typing back and forth.
To protect user privacy, SL2T sees sign language as a sequence of pose landmark locations rather than a raw camera feed. An on-device model (MediaPipe Holistic) tracks the location of points on the signer, and only these geometric coordinates are sent to the server for translation, allowing the original video to be discarded immediately.
The model behind the feature
SL2T was trained on more than 100,000 hours of signing data spanning over 50 sign languages, with roughly a quarter of the data in ASL. On the FLEURS-ASL benchmark, SL2T scores 70 BLEURT, well above previously reported results, according to Google.
Unlike spoken languages, which map sequential sounds directly to text, sign language translation presents unique computational challenges. Sign languages are independent linguistic systems with their own grammars and vocabularies, requiring true machine translation rather than direct word-for-word conversion.
AI’s ability to process spoken languages has advanced rapidly over recent decades, enabling automatic translation, dictation, and conversational interfaces that feel effortless to hearing users. Yet this technological revolution has not reached the world’s more than 200 sign languages, and the estimated 70 million Deaf and hard-of-hearing people who use them. Google says SL2T is the first sign language AI to ship in a real consumer product.
Google announced the feature on August 12, 2026. The model reaches consumers on August 20, 2026, when the Pixel 11 ships. The company is not charging extra for the feature. The Pixel 11 itself, which carries Google’s Tensor G6 chip, launched earlier this month.
Limitations Google discloses
The company acknowledges occasional errors with the technology, around rare signs, rapid fingerspelling, passive grammatical constructions, classifier depictions, and tense inference without sufficient context. ASL-to-English is the only language option at launch, and Google says the tool should not replace qualified interpreters in high-stakes settings.
The model also shows residual hallucination behavior, generating text when no one is signing, such as when a second person enters the frame or the signer pauses. Google says the release version significantly reduced these errors compared with pre-release builds that Deaf evaluators tested in July 2026, but has not eliminated them. Additional challenges include left-handed signers and one-handed signing, which becomes necessary when the other hand is holding the phone.
Deaf community involvement and governance
Google DeepMind states that Deaf perspectives shaped every stage of the project, from conceptualization by Sam Sepah, a Deaf Googler, to data collection with Deaf partners and evaluation in Deaf user studies.
The release carries an unusual layer of governance. Google DeepMind convened an advisory body, the AI Sign Language Advisory Committee, bringing together Deaf organizations including the National Association of the Deaf, the World Federation of the Deaf, the Deaf Professional Arts Network, and Rochester Institute of Technology’s National Technical Institute for the Deaf.
The AISLAC team also co-authored a public impact report covering SL2T 1.0’s capabilities and known limitations, an accountability mechanism that goes beyond a typical product launch.
Google DeepMind says more devices are coming soon, along with support for additional sign languages beyond ASL, though no timeline has been set. The approach echoes the broader push in the software industry toward building accessibility features alongside the communities that depend on them, rather than retrofitting them afterward.
Developers interested in building on the underlying architecture can follow Google DeepMind’s announcement for updates on a potential API and open-weights release under the SignGemma name that the company says is under active discussion.
The move also underlines how on-device AI pipelines are redefining what smartphone software can do without requiring a camera feed to leave the device, a privacy architecture increasingly relevant as AI demand continues to reshape the chip and device markets driving these capabilities.
