UMaine faculty investigating best uses for AI in special education

As the use of artificial intelligence and machine learning grows, so too do the questions about the best and most ethical way to deploy these technologies in various fields. In K–12 schools, for example, students, teachers, parents and other stakeholders face significant decisions about how to balance the potential benefits of generative AI and other tools with the risks, such as cheating, plagiarism and bias.

University of Maine associate professor of special education Sarah Howorth is addressing some of these questions through her research. Working with colleagues at UMaine and across the country, Howorth is seeking to identify best practices for using AI to support students with disabilities, as well as both the practicing and preservice teachers who instruct them, in inclusive settings.

“Many of these tools, including AI for text/language generation and speech recognition, could have major benefits for exceptional students, their teachers and families,” says Howorth, who is coordinator of special education graduate programs at the UMaine College of Education and Human Development and director of Maine Access to Inclusive Education Resources, a partnership between the university and the Maine Department of Education that acts as an information clearinghouse for educators, families and other professionals serving students with disabilities. 

“Some educators are already making innovative use of AI to design and deliver courses,” she says. “But as a society, we need to identify what it means to be smart when it comes to the use of these technologies in the classroom. How can they be deployed to strengthen uniquely human skills?”

Howorth is principal investigator of a project titled “Leading the Way: AI in Special Education Teacher Education,” which was recently awarded a $9,000 grant from the Council for Exceptional Children (CEC), the largest international professional organization dedicated to improving the educational success of children and youth with disabilities, as well as special gifts and talents. A longtime member and past-president of the CEC Division on Innovations in Special Education Technology (ISET), Howorth is working with experts from the CEC’s Teacher Education Division (TED) and Talented and Gifted (TAG) division on the project. 

The grant will allow researchers to collaborate across the three divisions to offer guidance and inform policy and pedagogical practices related to the rapid development of AI in education. It will support four webinars later this year featuring experts on topics related to AI and its use in the classroom, as well as creation of an award competition for educators on “Creative Use of AI to Support Exceptional Students and Inclusive Educational Practices.” The competition will include $500 prizes for up to three teachers who demonstrate innovative use of AI to support evidence-based, inclusive practices for all students. Howorth and her colleagues from the CEC Teacher Education Division also plan to edit a special issue of the Journal of Special Educator Preparation on artificial intelligence in special education teacher preparation. 

“Whether it’s using AI as an assistive technology in the classroom, creating objectives for individualized education plans, or example syllabi statements on the use of AI, we aim to highlight the best and most innovative practices to support teacher education programs, teachers, families, schools and students in using these tools,” says Howorth.

In addition to the CEC-funded project, Howorth and fellow UMaine special education faculty members Sara Flanagan and Melissa Cuba are involved with a year-long project led by the Center for Innovation, Design and Digital Learning (CIDDL) at the University of Kansas. They are taking part in a special initiative called the CIDDL Tech Alliance that is bringing together higher education faculty in the areas of special education and early intervention, as well as related service providers to increase the effective adoption and use of technology to support students with disabilities, specifically within personnel preparation programs. 

Howorth, Flanagan and Cuba participated in a recent webinar hosted by CIDDL Tech Alliance, where faculty members from participating institutions shared their plans to enhance technology integration in their special education preparation programs.

In one of her courses, Flanagan teaches students to use prompt engineering — structuring text so that it can be interpreted and understood by a generative AI model — to create materials that meet teachers’ instructional needs and students’ goals. For example, Flanagan says some students used OpenAI’s ChatGPT to reword a passage of text into a different reading level. They have also used it to write a social story, a type of narrative that educators and other service providers use to illustrate social norms to children with autism, as well as teach them and how to communicate appropriately with others or manage certain situations or problems. Other students used AI image generation to create images to pair with vocabulary words or other text. 

“When working with generative AI, we emphasize the importance of foregrounding teachers’ pedagogy and content-area knowledge, as well as their students’ instructional needs,” Flanagan says. “AI might be able to create any number of materials and resources, but teachers need to know how it supports their students and their instruction, and any changes or modifications that might need to be made to the generated content. AI is a tool, not a replacement for instruction.”

Flanagan is exploring how her graduate students, many of whom are practicing teachers, use generative AI and hopes to publish her research findings in the future.

Besides UMaine, the CIDDL Tech Alliance includes California State University, Chico; Georgia Southern University; the University of Hawaii at Manoa; Indiana State University; the University of Southern Mississippi; the University of Missouri; and the Washington Education Teacher Residency Program.

Contact: Casey Kelly,