American Sign Language (ASL) ranks as the third most widely used language in the United States, yet there is a significant lack of AI tools developed with ASL data compared to the abundant resources for the nation’s predominant languages, English and Spanish. To bridge this gap, NVIDIA, the American Society for Deaf Children, and the creative agency Hello Monday have partnered to create Signs, an interactive web platform designed to facilitate ASL learning and the formation of accessible AI applications. The platform offers users a validated library of ASL signs, allowing language learners to enhance their vocabulary with the assistance of a 3D avatar that illustrates each sign. Additionally, it features an AI tool that analyzes webcam recordings, providing real-time feedback on users' signing skills. Individuals at any proficiency level can contribute by signing particular words, aiding in building an open-source video dataset for ASL. NVIDIA aims to expand this dataset to 400, 000 video clips showcasing 1, 000 signed words, with fluent ASL users and interpreters validating each sign to guarantee accuracy. This will create a high-quality visual dictionary and teaching resource. “Most deaf children are born to hearing parents. Providing accessible tools like Signs for family members to start learning ASL early permits them to establish effective communication with children as young as six to eight months, ” commented Cheri Dowling, executive director of the American Society for Deaf Children. “With all vocabulary on the platform validated by professional ASL educators, users can trust what they are learning. ” NVIDIA intends to leverage this dataset to further advance AI applications that dismantle communication barriers between deaf and hearing communities. The data will be made publicly available as a resource for developing inclusive technologies, such as AI agents, digital human applications, and video conferencing tools.

It may also be utilized to enhance the Signs platform, providing real-time, AI-driven support and feedback for ASL applications throughout the ecosystem. **Enhancing ASL Education and Addressing Language Nuance** During its data collection phase, Signs already serves as a robust platform for ASL language acquisition, granting users the chance to learn and practice an essential set of 100 signs that improve their ability to communicate with friends or family members who utilize ASL. “The Signs learning platform could aid families with deaf children in quickly searching for specific words and observing the associated signs. It’s a practical tool to enhance their everyday use of ASL outside of formal classes, ” Dowling stated. “I envision both kids and parents exploring it together and having fun. ” While Signs currently emphasizes hand movements and finger placements for each sign, ASL also includes facial expressions and head movements that convey meaning. The team behind Signs is looking into how these non-manual signals can be observed and integrated into future iterations of the platform. They are also examining methods to incorporate other subtleties, such as regional variations and slang, into Signs, thereby enriching its ASL database. Collaborating with researchers at the Rochester Institute of Technology’s Center for Accessibility and Inclusion Research, they aim to evaluate and enhance the user experience of the Signs platform for deaf and hard-of-hearing users. “Enhancing ASL accessibility is a continuous endeavor, ” said Anders Jessen, a founding partner of Hello Monday/DEPT, which developed the Signs web platform and previously collaborated with the American Society for Deaf Children on Fingerspelling. xyz, an app that teaches the ASL alphabet. “Signs is designed to meet the demand for advanced AI tools that help bridge communication gaps between deaf and hearing communities. ” The dataset supporting Signs is expected to be released later this year.