Google is bringing sign language recognition to a mainstream smartphone for the first time, embedding a new AI model called SL2T into the Pixel 11 lineup that lets users sign in front of the camera instead of typing.<\/p>\n
The feature, announced by Google DeepMind on August 12, will be available when the Pixel 11 ships on August 20. It works inside Gboard and Live Transcribe, converting American Sign Language (ASL) into English text in real time. The launch marks the first time a sign language AI has shipped in a consumer product at scale, according to Google.<\/p>\n
“There is great diversity among deaf people in terms of their level of proficiency in signing, speaking, reading, and writing, so it is important to support access in all modalities,” DeepMind said in a statement. “Deaf people can benefit from sign language processing in the same way that hearing people benefit from spoken language processing.”<\/p>\n
How SL2T Works<\/p>\n
SL2T represents a departure from earlier sign language systems that relied on intermediate representations known as glosses \u2014 simplified labels for individual signs. DeepMind argues that glosses fail to capture the rich, non-linear aspects of sign languages, including non-manual markers like facial expressions and spatial grammar that are essential to ASL.<\/p>\n
Instead, SL2T translates directly from body landmark data to text. An on-device computer vision component called MediaPipe Holistic maps the signer’s face, hands, arms, and torso into geometric coordinates. Only those coordinates \u2014 not raw video \u2014 are sent to Google’s servers for translation.<\/p>\n
This architecture serves two purposes. It protects privacy by keeping actual footage on the device, and it reduces latency by transmitting lightweight coordinate data rather than video streams. DeepMind said the pipeline is optimized to minimize streaming delays, prevent hallucinations when the camera captures non-signing movement, and improve performance for one-handed signing while holding a phone.<\/p>\n
The model was trained on more than 100,000 hours of data spanning over 50 sign languages, with approximately one quarter of the corpus dedicated to ASL. DeepMind reported a BLEURT score of 70 on the FLEURS-ASL benchmark, which the company describes as state-of-the-art.<\/p>\n
Why Sign Language Has Lagged<\/p>\n
DeepMind frames sign languages as full natural languages with their own grammar and vocabulary \u2014 not simply “English on the hands,” as the company put it. That distinction explains why sign language AI has lagged behind spoken language processing for years.<\/p>\n
Two technical challenges have historically stymied progress. First, sign languages require translation rather than transcription, because their grammatical structures differ fundamentally from spoken languages. Second, signers convey meaning simultaneously through their hands, arms, torso, head, and face \u2014 a demanding computer vision problem that requires accurate tracking at high frame rates.<\/p>\n
Earlier efforts, including sign-language gloves, failed in part because they read hand shapes alone and ignored the rest of the body. SL2T was designed to handle both the visual perception and the translation in a unified pipeline.<\/p>\n
What Users Can Do<\/p>\n
On the Pixel 11, users can sign into Gboard anywhere they would normally type \u2014 searching the web, drafting messages, editing documents, or giving tasks to Gemini. In Live Transcribe, a user can sign a reply during a face-to-face conversation rather than tapping one out on the keyboard.<\/p>\n
Testers described signing in ASL as faster, more natural, and more comfortable than typing in English, according to DeepMind. The feature is available at no additional cost.<\/p>\n
At launch, SL2T supports only ASL-to-English. Google said it plans to roll out additional sign languages and expand to more devices over time. The company has also established an AI Sign Language Advisory Committee with Deaf organizations and sign language experts to guide deployment, and that committee will co-author a report on the model’s current capabilities and limitations.<\/p>\n
Globally, an estimated 70 million deaf and hard of hearing people communicate in roughly 200 sign languages. DeepMind acknowledged that one language pair is only a starting point, but noted that training SL2T on multiple languages simultaneously allows the model to learn shared underlying structures that could accelerate future expansion.<\/p>\n
The Pixel 11 Context<\/p>\n
The SL2T launch is part of a broader Pixel 11 hardware announcement that includes the Pixel 11, Pixel 11 Pro, Pixel 11 Pro XL, and Pixel 11 Pro Fold. All models feature Google’s new Tensor G6 chip, which the company says delivers a 50 percent improvement in TPU performance, enabling on-device AI processing up to 3.5 times faster while consuming up to 3.5 times less energy.<\/p>\n
The Pro models introduce HiLight, a ring of RGB LEDs around the rear camera flash that lights up when Gemini is listening or processing, and can alert users to calls from specific contacts with customizable colors. The camera bar has been slimmed down, with Google claiming a 40 percent reduction in camera thickness on some models.<\/p>\n
Pricing starts at $899 for the Pixel 11, up $100 from last year, and reaches $1,899 for the Pixel 11 Pro Fold. Preorders opened on August 12.<\/p>\n
ModelStarting PriceKey FeaturePixel 11$899Tensor G6, SL2T sign inputPixel 11 Pro$1,099HiLight LED ring, 120x zoomPixel 11 Pro XL$1,199HiLight LED ring, larger displayPixel 11 Pro Fold$1,8998-inch foldable display, 16GB RAM<\/p>\n
Note: Prices reflect Google Store US listings announced August 12, 2026. Pro and Pro XL pricing based on reported $100 increase across the lineup.<\/p>\n
The sign language feature arrives during a period of leadership transition at DeepMind. In early August, Demis Hassabis moved from CEO to chairman, with Koray Kavukcuoglu taking over day-to-day operations. The Gemini app has surpassed 950 million monthly users.<\/p>\n
For the deaf and hard of hearing community, SL2T represents a meaningful step toward parity with the voice dictation that hearing users have enjoyed for years. For Google, it demonstrates how the company’s AI research can translate into concrete accessibility features that differentiate its hardware in a crowded smartphone market.<\/p>\n
The next items on DeepMind’s roadmap include adding more sign languages and developing models that can generate sign language output, rather than only reading it \u2014 a capability that could eventually enable two-way communication between signing and non-signing users.<\/p>\n","protected":false},"excerpt":{"rendered":"Google is bringing sign language recognition to a mainstream smartphone for the first time, embedding a new AI…\n","protected":false},"author":2,"featured_media":138242,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[9],"tags":[68288,68286,5044,23265,2408,132,7543,68285,68287,25479,68284,60145],"class_list":["post-138241","post","type-post","status-publish","format-standard","has-post-thumbnail","category-google","tag-ai-sign-language-advisory-committee","tag-american-sign-language","tag-deepmind","tag-gboard","tag-gemini","tag-google","tag-google-deepmind","tag-live-transcribe","tag-mediapipe-holistic","tag-pixel-11","tag-sl2t","tag-tensor-g6"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/138241","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=138241"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/138241\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/138242"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=138241"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=138241"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=138241"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}