AI for AAC: How is AI used to improve high-tech AAC?

As someone who loves conversation and humor, Jordi was devastated when he found out that he was going to lose his ability to speak. After his ALS diagnosis, he struggled to participate in social situations as his voice began to weaken.  

Fortunately for Jordi and many others, AI was advancing to the point where it could be used to create digital copies of voices. Eventually, he was able to work with our voice partner Acapela to create a clone of his voice. Since pairing it with a speech device, he hasn’t stopped talking. 

He even started making videos about the realities of living with ALS. Using his digital voice to entertain and inform his audience, Jordi has gained over 100,000 followers on various social media platforms. 

Creating natural-sounding synthetic voices is just one application of AI in assistive communication. Here are four ways AI can enhance high-tech AAC. 

Synthetic voices (Deep learning) 

To create natural, expressive, and individualized voices, we integrate cloud-based neural voice engines that use deep learning. Next-generation synthetic voices will give users the most human sound we've ever offered and will include even more languages. 

TD Snap (Deep neural networks) 

Deep neural networks in TD Snap allow for accurate grammar prediction. Intelligent language capabilities make communication more efficient by automatically adjusting word forms such as tense and plurals. Deep neural networks also reduce the number of steps required to create messages. 

TD Control & TD Browse (Convolutional Neural Networks) 

Convolutional neural networks in TD Control and TD Browse allow for greater vision system accuracy, giving users more reliable gaze access. This form of AI interprets gaze input more accurately improves selection accuracy through advanced pattern recognition and supports users with complex movement and positioning needs. 

Eye tracking (Machine learning) 

Our eye trackers have been powered by machine learning for over a decade. Machine learning helps them understand and respond to each user by improving performance through personalized calibration, estimating eye geometry and tracking their gaze accurately, adapting to lighting, posture, and individual differences. 

Ongoing explorations of AI for assistive communication 

We are currently exploring the uses of generative AI in assistive communication. As we experiment with different applications, we keep customer experience at the forefront of our decisions. Our goal in incorporating AI technologies is always creating more comfortable and personalized communication for people with disabilities. 

We’ll stay engaged in conversations across the assistive technology industry about where AI adds value and where a more thoughtful approach is needed. Advances in AI must address specific needs within the disability community; we intend to develop products and features that people actually want to use. 

Learn about how we use AI responsibly