Summer interns harness AI for accessibility

Aug 04, 2025

Gallaudet senior Austin Baker stands in front of a video with an image that looks like a child has played dot-to-dot on. As part of his summer internship, he’s trying to see if AI can be used to identify American Sign Language use in a video. “There are many people with disabilities in the world who have different kinds of access issues that could be resolved through technology,” he says.

Baker is one of five interns participating in this summer’s AICT REU undergrad program, which specifically focuses on inclusive information and technology through design and accessibility research. After ten weeks of conducting AI-related research and analyzing results, students from Gallaudet, the Rochester Institute of Technology (RIT), and Cornell University wrote papers and presented posters on their findings at an event on July 24.

Principal investigator of the REU grant, Dr. Raja Kushalnagar explains the AI focus. “In this age of generative AI and sign recognition, we can’t permit technology to pass by without involving Deaf and hard-of-hearing people. Our perspectives are essential,” he says.

Funded by the National Science Foundation, the REU mechanism provides engaging research opportunities for undergrads to support developing careers in STEM research. Christian Vogler, Director of the Technology Access Program, is the Co-Principal Investigator on the grant. A third faculty mentor, James Waller, supported the interns, along with Graduate research assistants Amy Luna, Joe Merino, and Paige DeVries, ’22 & G-’24. This year marks the 10th cohort participating in the AICT REU program.

Gallaudet faculty have secured more grant funding in the pipeline for accessibility and AI. “People are already using AI technologies for language access and in trying to obtain equity between spoken and signed language. It’s time for us to capitalize on that momentum,” says Kushalnagar.

Making tech more accessible with AI

Read on to learn about the interns’ projects this summer.

Senior Austin Baker’s project evaluates Google’s MediaPipe to identify when ASL is occurring in a video, finding it did well in high-quality video samples, but not so much in archival footage.

Aside from the highly technical projects they grapple with, Baker, who is majoring in IT and minoring in Data Science, says the internship also provides him with a range of career skills. “This is real-world experience. As an intern, I have to work well with other people, and I have a deadline to complete tasks. In school, there’s some flexibility–the teacher might extend a deadline, but that’s not an option here. I have to finish things right on time,” he says.

Gallaudet students Sarah Alghamdi and Isaac Benjamin from RIT evaluate whether a Language Learning Model improves communication between hearing people and those who are Deaf or hard of hearing on Zoom.

They created an AI bot to constantly analyze the transcripts of Zoom meetings and suggest responses and prompts for chat users.

Sabrina Boyle from RIT researched how to improve AI models used for ASL sign recognition.

Noticing that sign language recognition efforts are usually filmed with the signer straight on, Boyle’s team set up a panoptic studio using 360-degree videos to collect more realistic sign language data.

Cornellian Adriana Jimenez-Bonilla investigated methods for improving results when using English to ASL dictionaries and evaluated whether AI could help disambiguate the meaning and context of ASL signs.

Find out more about Gallaudet’s outstanding REU summer research programs: Accessible Information and Communication Technologies (AICT) and Cognitive and Educational Neuroscience Training and Research Experience (CENTRE)

Areas of Study

Summer interns harness AI for accessibility

Making tech more accessible with AI

Recent News