Student teams create AI tools to help neurodivergent learners
Team Capstone, which participated in the Arizona AI Challenge, included undergraduate students (from left) Chandler Farrington, computer science; Stevie Cervantes, computer systems engineering; Journey Hancock, computer science; and Dvir Hamu, computer science. Photo by Edmundo Mendez/Dez Visuals
When Stevie Cervantes was a first-year student at Arizona State University, she shied away from using AI.
“AI was everywhere and my friends were asking me, ‘Oh, do you use ChatGPT?’ And I hadn’t even explored it because I was afraid I would be accused of cheating,” she said.
But this fall, Cervantes, now a fourth-year student majoring in computer systems engineering, dove into AI. She was among more than 90 students who entered the Arizona AI Challenge and her team is one of two finalists chosen to participate in the next level — the Global AI Challenge.
The students developed AI-enabled tools to help people with neurodiversity navigate their post-college lives.
The competition was sponsored by the Spark Center for Innovation in Learning, which was created earlier this year by former U.S. Sen. Kyrsten Sinema, who is a Distinguished Professor of Practice in the School of Social Work at ASU. The center is dedicated to harnessing AI to help neurodivergent learners.
“With the Spark Challenge, I saw how AI can be applicable to everyday life and be useful, and it’s definitely changed the way I see it,” Cervantes said.
The Arizona AI Challenge was open to all college students in Arizona and the 20 teams included students from ASU and Pima and South Mountain community colleges.
Ten teams were named semi-finalists and participated in a five-day hackathon in November to prototype AI tools to help people with autism, dyslexia, ADHD or mental health conditions succeed after college.
The two finalist teams each won $5,000 and are now fine tuning their proposals for the global competition.
Solutions that hit close to home
Students with autism, attention deficit disorder or dyslexia have support while in college, but often those supports disappear when they graduate and begin careers. The AZ AI Competition task was to develop an AI-enabled tool that can help neurodiverse people with mental skills such as starting and completing tasks.
That real-world problem-solving is what drew Jacob Kuriakose and his team, Vertex AI, the other finalists.
“Many of us have friends or classmates who are neurodivergent and rely heavily on campus accommodations to stay organized and succeed. And the moment they graduate, all of that disappears,” he said.
“So when we read the challenge description, it felt like an opportunity to build something truly meaningful. The motivation was both personal and technical.”
The five-day timing was intense.
“There were many moments where something broke at 2 a.m. or a feature stopped working right before we needed to record our final demo,” said Kuriakose, a graduate student majoring in data science who was the AI architect on his team.
“Over the last two days we were awake the whole night working on this project. But I think that is what brought us closer and made the win meaningful and impactful.”
Vertex AI’s product is Navia, an AI assistant that breaks tasks down to micro-actions (“gather ingredients,” “prepare ingredients,” “cook the meal”) and offers a Tinder-style peer-matching network to build community.
Cervantes’ team is called Capstone because the four members met while working on their engineering capstone projects. She worked on the front end, creating the overall design and user experience.
“I was doing everything online, but for those five days, I'm pretty sure we didn't do anything but work on that project,” she said.
“And then the technical problems started to kick in. We were hitting roadblocks with different (application programming interfaces) and different libraries and trying to figure out a framework.
“It was definitely a learning process for all of us because a lot of us have never worked with AI in that level.”
Team Capstone created an AI assistant called Nerva, which helps users break down tasks into small chunks and then schedules each task in a calendar, offering supportive feedback.
Interactive journaling personalizes Nerva. At the pitch demonstration, Nerva asked, “Would you like me to play some Beatles to help you focus while you start?”
The team also created a physical version of Nerva, like an Amazon Echo, to sit on a desktop.
Both Vertex AI and Capstone are refining their proposals for the Global AI Challenge and the chance to win the Sinema Student Prize, a $15,000 seed fund. The four finalist teams for the global contest will pitch at the ASU+GSV Summit in April.
The competitions are also sponsored by the J. Orin Edson Entrepreneurship + Innovation Institute at ASU, whose executive director, Linda Ricchiuti, was a judge for the student contest.
All of the student teams who entered the Arizona competition can continue in the Edson Institute, getting mentoring support and pitching for funding to develop their projects.
“The biggest non-technical takeaway is the entrepreneurial aspect, which is so fascinating because we don't learn that as engineers,” Cervantes said.
“I’m really excited to see where this product goes.”
Growing potential
Sinema, who spoke at the “Agentic AI and the Student Experience” conference in October, told the crowd that she was inspired to create the Spark Center for Innovation in Learning at ASU because of her early-career work as an elementary school social worker.
“I saw teachers in overcrowded classrooms with not enough of the tools and skills they needed to meet the very specific educational needs of every student. And I don't mean that as a knock on the education system or a knock on the school even, because it turns out there is no way to actually have all the skills and tools that you need to meet the specific learning needs of every student in your classroom.
“Instead of, ‘We want to figure out how to have you behave better in the classroom or get you to do the worksheet,’ … I want to figure out how we help these kids grow to be everything that their potential allows them to be.”
AI can transform the way that neurodiverse learners are identified, she said.
“You could use AI as a tool to identify a student who has dyslexia very early on, long before reading even begins. Or help students who may be on the spectrum use AI to interact in a way that is most appropriate for them to communicate,” she said.
She said that OpenAI and Microsoft will help the competitors develop their projects.
“We are going to help turn those good ideas into reality so they go into the world and help real people,” she said.
More Science and technology
Findings on adenoviruses in baby gelada monkeys provide a window into our own cold and flu season
If you have young kids or spend time around day care centers, you know the drill: Someone gets a cold, and soon the whole group is sniffling and sneezing. Now imagine that same pattern playing out in…
ASU researchers shed light on ancient tattoos in the Nile Valley
Long after an archaeological excavation, discoveries can still be made. One such example of this is newly discovered tattoos from ancient Nubia, nearly doubling the amount previously known from the…
ASU’s LEAPS lab marks a decade of energy impact
Nathan Johnson doesn't mince words when it comes to Earth's energy requirements.“The world needs every electron it can get to match the accelerated pace of economic development and electrification,”…