Bringing Sci-Fi Future to Life

Chancellor Woolsey grew up in a small Missouri town where he spent his plentiful free time marveling at fictional depictions of the future. He watched computers advance, and in his college years 3D printing, then ChatGPT. Somewhere between Star Wars marathons and a lifelong love of science fiction, he developed a question that would follow him to a PhD program at the University of Arizona: what if the technology we've been dreaming about on screen could actually help people? 

Now a third-year doctoral student in the Management Information Systems department, Chancellor is turning that question into something real in the Deep Target NLP Lab, funded by the National Institutes of Health (NIH) and People Acting Now Discover Answers (PANDAS). His projects include two AI-powered tools designed to address gaps in health care for children. 

When Chancellor was looking at various tools, he found applications filmmakers and game designers use to convert videos into avatars. He thought, “I work a lot in the medical side, so this would be a great way to create HIPAA-compliant video data.” 

Chancellor and a team of autism researchers used these tools on a solution for screening children for autism in a way that's private, accurate, and scalable. The VAST pipeline converts video footage into an anonymized avatar, then further reduces it to a "skeleton tensor," a compact mathematical representation of a person's joint movements over time. The result strips away everything a machine doesn't need—background, clothing, facial features—and keeps only what matters to it: motion. "We've given these machines a very simplified view of the world," Chancellor says. "Which is very useful for automation." 

The AI answering a binary question of, “Autism, yes or no?” on the tensor data is extraordinarily fast and accurate—99%. Even better, reducing video files to tensor skeletons creates smaller file sizes that will be less burdensome in low-resource settings like rural health care or a general pediatrician’s office. It is not meant to be a standalone diagnostic tool, but it is another tool a doctor can use in an evaluation.  

The project Chancellor is now working on is JESSE (Just-In Time Empathetic Speech and Support Interaction Engine), a true interdisciplinary partnership and collaboration between AI developers, researchers, and clinical experts (pediatricians and speech language therapists). It is an AI-driven avatar that serves as an on-demand practice partner for children working through therapeutic exercises. The problem it addresses is straightforward but stubborn: schools dealing with students with special needs often have few or overextended resources. Therapists are stretched thin, and keep business hours, while kids (and parents) have a hard time fitting sessions into their day. 

JESSE meets the children where and when they are. Whether a child wants to practice asking for help, initiating conversation, or simply talking about their day, the avatar is available whenever they are, by voice or text, and logs each session for the supervising therapist to review. "We're not cutting therapists out of the loop," Woolsey clarifies. "This is a helper tool, not a replacement tool. 

Chancellor likes avatars as a way to increase human-AI communication and collaboration, and help decrease aversion to AI. He says, “Right now, I think one of the limitations of ChatGPT or these other AI models is you have nothing to stare at. It's a man behind the curtain.” He also sees that avatars could be helpful when someone thinks they might be having a mental health issue. “With AI, can we do this in a way that allows us to not even have to have a human in the room, because humans are scary.” Although currently, people can sometimes see AI as scary.  

Chancellor’s work is focused towards building AIs that help people and getting others to see AI as helpful. “How can we break past the scary barrier, back into: oh, hey, I love seeing this in Star Trek, Star Wars.” He is still very much the sci-fi nerd—just one trying to build Baymax, the healthcare robot from Big Hero 6, before anyone else does.