Social Work faculty recognized as Warrior Trailblazer for groundbreaking AI research innovations
Bryan Victor, an associate professor at Wayne State University’s School of Social Work, received the 2023 Wayne State University Warrior Trailblazers Award.
The Warrior Trailblazers Award is one of the university’s six Value Awards, celebrating individuals who significantly impact others and contribute to the university’s mission of creating and advancing knowledge, preparing a diverse student body to thrive, and positively influencing local and global communities. This award specifically honors employees who drive outstanding innovation that benefits the Wayne State community by demonstrating the courage to implement new initiatives that enhance processes and improve the customer experience. Victor was honored with the award, presented by the university’s Human Resources department during the annual Employee Recognition Week Ceremony in April 2024.
Victor was nominated by the School of Social Work Acting Dean and Professor Debra Patterson in recognition of his groundbreaking research innovations with artificial intelligence, particularly his use of large language models like ChatGPT in social work scholarship. Acting Dean Patterson noted, “I nominated Dr. Bryan Victor for the Warrior Trailblazer Award for his innovation in using generative AI to advance the social work profession. He has brought prominence to Wayne State University and the School of Social Work by successfully applying new scientific methods to longstanding social problems. Using large-language models in his scholarship has major methodological implications and provides practical, real-world impact that benefits children and families.”
The rapid advances in machine learning and large language model technology have revolutionized how Victor and his colleagues approach data classification and coding. For instance, the child welfare system has vast amounts of unstructured text data from files, case notes, and other documentation. Extracting valuable insights from these records has traditionally been extremely time-consuming, often taking months to write the necessary computer code and clean the data for analysis. However, with the use of large language models, there is no longer a need for data cleaning, and their capacity and accuracy are remarkable.
“So many agencies are sitting on troves of information that might help them make more data-informed decisions about how they serve their clients and communities,” he said. “This new technology allows us to unlock information that we wouldn’t have had the capacity to in the past.”
Victor also led a series of studies that used ChatGPT to evaluate the Association of Social Work Boards (ASWB) licensing exams. A 2022 report identified severe racial disparities in pass rates on the ASWB exams. Using large language models, Victor and his colleagues identified several new validity concerns with the exams. That research has been influential to policy makers across the United States as they consider alternative pathways to social work licensure that correct for the negative impact and lack of empirical support for the current exams.
While Victor is excited about the potential of emerging technologies to advance social work practice and research, he emphasizes the importance of protecting individuals’ personal information. He and his colleagues are addressing ethical usage and data security concerns by utilizing smaller local models, which offer a more secure approach, and do not transmit sensitive data to third parties.
“I appreciate that the university has taken the time to acknowledge employees who are exploring innovative approach to their work,” Victor said. “I think the acknowledgement of AI is an important one, and it’s something that people in social work will really have to consider. These early explorations – thinking about safety and ethics – is a conversation that we need to be having in this field right now. The award was very unexpected, and I’m very grateful to have received it.”
Writer: Sarah Freeman and Editor: Betsy Vanderstelt