Google's Gboard Developing New Feature to Translate Sign Language to Text

Google has revealed that it is implementing a new feature called “Sign-to-text” in its Android keyboard app Gboard, which reads sign language and converts it to text. Android Authority analyzed the code within the Gboard beta version and discovered the unreleased feature. Android Police reports this.

This feature recognizes the user’s sign language through the camera and uses AI to convert it into text. Google has added a privacy notice, explaining that while the video itself is not sent to the cloud, the sign language data points may be sent to the cloud for processing.

Discovery Background and Current Status

Android Authority analyzed the Gboard beta code and confirmed the existence of this feature. However, it was not possible to actually make the feature work, and only user interface images were obtained. The interface images show that when the feature is launched, a message such as “When you use sign language toward the phone, it will be converted to text with the power of AI” is displayed.

At this point, it is merely a feature existing in the code, and a long road is expected before a public release. It is unclear whether Google will ultimately include this feature in the final product. In the past, Google has developed accessibility features but canceled them before release.

Language Support Challenges

Sign language has different linguistic systems depending on the country or region. There are diverse sign languages around the world, such as American Sign Language (ASL), British Sign Language (BSL), and Japanese Sign Language (JSL). Which sign language Google will support is a major factor influencing the spread of this feature.

Based on available information, it is likely that the initial release will be limited to specific regions. On the other hand, if Google has sufficient training data, the possibility of a global rollout is not zero.

Impact on Accessibility

The ability to convert sign language to text has the potential to expand communication methods for the hearing impaired. Especially in communication with people who do not understand sign language, real-time text conversion could be an important tool for removing barriers.

Gboard already supports voice input, and in addition to typing and swipe input, adding sign language as a new input method. This effort is seen as part of Google’s accessibility strategy, but the balance between accuracy, supported languages, and privacy protection will be key to practical use.

Editorial Opinion

In the short term, the focus is on whether this feature will transition from beta to the final product. Google has several cases in the past where it experimented with accessibility features but never released them. It should be recognized that the existence of code alone does not guarantee productization. Also, whether sign language recognition accuracy has reached a practical level will become clear in future updates.

In the long term, the establishment of sign language recognition technology could bring about a paradigm shift in communication methods. If sign language can be treated as text, interactions with chatbots and voice assistants will be opened up to the hearing impaired. However, the fundamental challenge remains of how AI will learn the regional differences and cultural backgrounds of sign language.

What the editorial department wants to ask is whether this feature truly improves accessibility or is merely part of Google’s data collection strategy. The explanation in the privacy policy that “the video is not sent to the cloud, but data points are sent” is potentially misleading for general users. There is a need for clarification on the anonymization and deletion policy of sign language data.

References

• “Gboard’s next big feature could translate your sign language into text”, by Timi Cantisano — Android Police, 2026-07-17T22:01:33.000Z (ARR)
• Source URL: https://www.androidpolice.com/gboard-sign-to-text-in-the-works/