Nvidia

Signs: An AI-powered platform for learning sign language

Transcript

Imagine being unable to tell your child, I love you. In the US 11 million people are deaf or have significant difficulty hearing. 90% of deaf children are born to hearing parents who can’t communicate with them. As a hearing mom, having a deaf-hearted hearing child, I just feel left out communicating with her. And it hurts me. Studies show that the earlier a child is exposed to sign language, the better their language development is overall, but it also impacts their cognitive development, their mental health, and overall educational success. It ripples out throughout their life. It’s that important, and parents are critical to that language learning because I want to talk to her about her wedding, her grandkids, you know, buying a house or being a lawyer or a doctor. I wanna be able to be in her life and I know that in order for me to do that, I have to learn that language and I have to learn that culture. That’s why we partnered with Nvidia, the world leader in AI computing, and the American Society for Deaf Children to build Signs, an AI-powered platform that revolutionizes sign language learning by blending advanced technology with an intuitive user-friendly approach through your camera and an interactive avatar. The platform provides step-by-step visual guidance, tracks your movements in real time, and offers instant feedback. This empowers users to learn sign language in a way that’s never been possible before. But Signs is more than a learning tool. It’s a resource for the future of ASL. This is a really wonderful system, could improve the lives of the deaf and hard of hearing community right away. Opportunities are just endless, and this is just the beginning. The more people who use Signs, the better it gets.

(  Services  )

  • AI
  • Customer Experience
  • Tech & Data

The challenge

For 11 million Deaf and hard-of-hearing individuals in the U.S., access to effective communication is a daily challenge. And for Deaf children, the stakes are even higher. 90% are born to hearing parents who may have little to no experience with sign language. Without early exposure to American Sign Language (ASL), these children risk language deprivation—a condition that can impact cognitive development, education, and emotional well-being.

a father teaching his son sign language


But learning ASL on your own has always been difficult. While there are incredible teachers, ASL programs, and schools that offer instruction, many people don’t have access to them. Traditional ASL classes can be expensive and geographically limited. Video tutorials exist, but they lack the one thing learners need most: feedback. They don’t tell you when your hand position is slightly off. They don’t help you adjust your movements in real-time. They don’t respond to you.

ASL education needed a revolution—one that could make learning more interactive, more accessible, and more immediate. One that could empower learners, bridge the gap between the Deaf and hearing communities, and make mastering ASL more intuitive for anyone, anywhere.

signs app
purple robot hand

The solution – Signs

In collaboration with NVIDIA and the American Society for Deaf Children, we created Signs—a first-of-its-kind, AI-powered sign language platform that makes learning ASL as natural as a conversation.

Using cutting-edge AI, computer vision, and machine learning, Signs transforms any camera into an interactive sign language coach.

  • Step-by-step guidance: a 3D avatar demonstrates each sign from multiple angles
  • Real-time feedback: AI tracks the user’s hand movements, providing instant corrections
  • A growing ASL database: users contribute videos of themselves signing, helping to train the AI and build an ever-expanding, open-source dataset

For the first time, learners don’t just watch—they interact. They get immediate, personalised feedback. They practice with confidence, knowing that every correction brings them closer to fluency.

And it’s working.

screenshot of signs app

The impact

Signs launched with zero paid media—and still, the world listened.

  • Featured on CNN, Axios, Venturebeat, and across global media.
  • A rapidly expanding dataset fuelling the future of AI-powered sign language recognition.

+1b

earned impressions

+20m

people reached in the first week

+20,000

signs learned in 10 days

Building a global AI movement

Signs is not just teaching ASL—it’s shaping the future of AI-powered accessibility:

  • 400,000+ video clips and 1,000+ signs in development.
  • Continuous AI training, refining recognition of complex sign movements and expressions.
  • A dataset made publicly available by NVIDIA, allowing researchers and developers to build the next generation of AI-driven accessibility tools—from real-time sign language translation in video conferencing to an AI assistant that understands sign language.

By opening up this dataset, Signs isn’t just helping individual learners—it’s fuelling a future where AI doesn’t just recognise sign language, but understands it.

Signs is free, open to all, and constantly improving.

  • Learn – Start signing today at signs-ai.com
  • Contribute – Record and upload signs to help expand the dataset
  • Share – The more people use Signs, the smarter it becomes
“

Ultimately, our goal is to connect families, friends, and communities by making ASL learning more accessible, while simultaneously enabling the creation of more inclusive AI technologies.


”

Michael Boone, NVIDIA

robot saluting

Next project

INTER

Inspiring and entertaining fans all over the globe

View Work