Why are Black students more likely to be falsely accused of using AI to complete assignments?

When Taylor McGee, assistant professor at Christopher Newport University, first read the Common Sense Media study, “The Dawn of the AI Era,” one finding jumped out at her. Black students reported that they were about twice as likely to have their work incorrectly flagged as generated by artificial intelligence tools as their white and Latino peers. McGee, driven by her curiosity to understand just what was going on here—from the perspectives of the students and the teachers—designed a research project to investigate this finding. 

Her forthcoming study, “Authorship, Assumption, and Equity: How AI Shapes Teachers’ Perceptions of Black Student Work,” was recently awarded a grant from the Spencer Foundation. The two-year study will examine how current and preservice teachers think about and respond to students’ use of AI for schoolwork. For McGee, researching the potential academic and psychological harms to Black students that incidents like this cause is central to her why she entered the field of education in the first place. And the opportunity to influence educational practice and policy that could change how teachers interact with Black students is “the heart of why I’m doing this career,” she says.  

This is not a new phenomenon for Black students. “We know a lot about how Black students are penalized and pushed out of schools for things that are cultural expressions—Black girls receiving out-of-school suspensions for their choice of physical expression, including hairstyles and clothing, engaging in the cultural tradition of call-and-response,” McGee says. These types of punishments can not only impact academic outcomes but are also related to higher levels of unemployment, fewer job prospects, reduced salaries and the school-to-prison pipeline, McGee explains.

A seemingly small incident, like being falsely accused of using AI to complete assignments, can have a major impact on a student’s academic trajectory, McGee says. It can “put them off from wanting to demonstrate their knowledge and academic excellence,” she says. For the Black students affected by this, the stakes are high. 

McGee wants to know what is going on the students end, and the teachers end. Her preliminary conversations with teachers have shown her that in large part, everyone is unsure how to handle AI, “and being unsure can lead to negative experiences for students,” McGee says.  

And while the educators' implicit biases are likely playing a part in the false accusations the Common Sense Media study reported on, McGee also points out that these same biases are baked into AI-detection tools. She plans to examine these potential biases in her research, both at the educator level and in the AI-detection tools that they are using. 

“Bias is transmitted through educational experiences, and AI is becoming an increasingly central part of how those experiences are shaped,” says USC Rossier Professor of Research Kendrick Davis, who is a part of the study’s research team. “As these tools influence how students are judged, supported and afforded opportunity, we have to ensure—in both practice and policy—that they are not reproducing the very structural barriers education is supposed to dismantle.”

This project will be the first that McGee has led as a principal investigator. During her postdoctoral work at USC Rossier under Professor Brendesha Tynes, she also collaborated with Davis on several projects, including as a co-author of the three inaugural policy reports issued by USC’s Critical Policy Collective in 2025. The CPC has allowed her to explore how research implications from educational psychology studies can impact educational policy, McGee says. McGee is still involved with the collective and provides mentorship and leadership to the USC Rossier PhD students with whom she collaborates. 

As she planned the study, McGee tapped Professor of Research Kendrick Davis to assist with the project. His insight was instrumental in shaping the study’s logistics and budget, she says. McGee will use a mixed-methods approach in the study, an approach she credits to Professor of Education Jessica DeCuir Gunby. She will integrate both quantitative and qualitative components into the research, conducting teacher questionnaires and interviews while considering the larger factors at play. The study will collect data in both Virginia and Los Angeles, with Davis leading data collection in L.A.

Phase one will focus on studying the attitudes and approaches to AI of current and preservice teachers in both locations. Phase two will look at the process these teachers employ as they identify AI-generated content. In the third phase of the study, McGee and Davis will consider how implicit biases may be related to teachers’ evaluations. 

AI technology is outpacing schools’ ability to respond, yet it is already reshaping how students learn and teachers teach. Through this study, McGee aims to provide insights to schools and educators to help them navigate this new AI-powered landscape, while ensuring they are attentive not only to their own implicit biases but also to those embedded in the technology itself. 

“The national conversation around generative AI in schools often misses the human element—specifically, how these tools, and the distrust of students that comes with it, impact teacher-student dynamics and, crucially, how they can negatively affect student motivation and long-term outcomes,” McGee says. “This grant allows us to bridge that gap.”