Clemson awarded $3 million grant to develop AI-powered professional development for teachers

A teacher’s search for the best way to sharpen their skills in the classroom may soon be as easy as browsing Netflix for a new movie. Clemson University researchers have been awarded $3 million over three years to develop a personalized professional development recommender system for teachers that resembles the way businesses such as Amazon.com recommend products or streaming entertainment.

Researchers from the Clemson University College of Education and the College of Engineering, Computing and Applied Sciences will develop this recommender system to improve teacher effectiveness and retention while increasing student achievement. The grant award comes from the U.S. Department of Education Supporting Effective Educator Development Grant Program. Of 130 applications to the program, Clemson’s proposal is one of only 12 to be awarded.

Jeff Marshall serves as associate dean for research and graduate studies in the College of Education as well as principal investigator for the research project. He said the project seeks to offer a more effective alternative to the one-size-fits-all approach to professional development for teachers, which he said is often not appropriately responsive to the needs of teachers and schools.

“Professional development is a $2-3 billion a year industry, and it’s critical to support and sustain teachers,” Marshall said. “Online services offered by the business world have done an extraordinary job using artificial intelligence and recommender systems to inform how we live our lives. It only makes sense that we can apply those same successful strategies to the way we help our teachers develop in the classroom.”

Marshall said the issue is not a lack of quality professional development programs; it’s the way those programs are matched to the educator and vice versa. Instead of simply offering teachers a large library of positively rated courses or books and asking them to pick one, the researchers will seek to discover which particular programs or combination of programs begin to move teachers and students in successful directions.

The system works by first using in-person guidance in an initial set of schools. The data captured by researchers in these schools is then used to design the AI-powered recommender system, which can then be used with a larger audience of teachers. As the system expands to provide more professional development opportunities to more teachers, the researchers monitor its progress to ensure it continues making effective connections between educator and instruction.

According to Nathan McNeese, assistant professor in the Clemson University School of Computing, the recommender system and the user interface will both be human-centered, meaning real humans will motivate and have input on how the larger system works to recommend professional development.

McNeese said he was attracted to the project because of its relevancy and need in the South Carolina and beyond. He believes the project will help advance K-12 education by building something unique that can be utilized specifically by educators in a meaningful way.

“In my opinion, we need to be developing and producing more socio-technically relevant work in the education sector,” McNeese said. “This project is really the best of many worlds: it is motivated by real world educators, allows for interdisciplinary computational systems to be built, and will allow for contributions that are both conceptual and theoretical but also applied.”

Anne Pressley, director of the Office of Standards and Learning at South Carolina Department of Education, will serve on the project’s advisory board to oversee strategy and provide strategic advice to the research team. She said the concept that the research team is exploring and planning to implement is an interesting potential solution to issues surrounding professional development, teacher retention and student learning.

Researchers hope the recommender system will begin to prove its worth in multiple locations across the state. Although the initial data capturing will occur in pilot schools in the Upstate, researchers plan to partner with 12-17 districts across South Carolina. Marshall said this eventual expansion is crucial in order to benefit underserved communities across the state and ensure the research is broadly applicable.

George J. Petersen, founding dean of the College of Education, is involved in the research to ensure strong partnerships among Clemson, partner districts and individual schools. Petersen said the research’s potential impact in the areas of teacher retention and student learning align with the college and university’s land-grant mission to serve the people of South Carolina.

This project is designed to enhance the professional competence and confidence of teachers through the development of personally tailored professional development programs,” Petersen said. “The impact of this type of professional development on teachers’ knowledge and skills will also greatly benefit their students’ learning and experiences. This is just one more way the best-in-class College of Education serves all of South Carolina.”

END