SU professor hosts class on AI, ethical image classification
Get the latest Syracuse news delivered right to your inbox.
Subscribe to our newsletter here.
Syracuse University Professor Jennifer Stromer-Galley hosted a Zoom sample class titled “Data Justice: Navigating Ethical Image Classification in the Age of AI” on Wednesday to educate prospective SU undergraduate students about artificial intelligence.
Stromer-Galley teaches Information Studies 343: Data and Society and is the senior associate dean for academic and faculty affairs for SU’s School of Information Studies. The class was held for prospective students to learn the history of image classification technologies and their problems, Stromer-Galley said.
The iSchool’s Director of Admissions and Recruitment Zachary Schuster, who joined Stromer-Galley in the class, said it’s important for students to hold a basic understanding of AI as it advances.
“Understanding their benefits and implications helps students make informed decisions, contribute to responsible development, and prepare for technology-driven careers,” Schuster wrote in an email statement to The Daily Orange.
Stromer-Galley discussed the biases that artificial intelligence places on faces and how that amplifies stereotypes. She first showed Google search results of both a young and old professor to ask students what they saw; their answers consisted mostly of white men. She then moved to describe image classifiers, which she described as AI tools that distinguish and label differences between different images.
One image classification tool is ImageNet, an AI database which places images into hierarchies. When Stromer-Galley searched images on the database in the class, words associated with the pictures generated on the screen, which she recognized as problematic for associating stereotypical labels to faces.
With ChatGPT and similar AI tools emerging, Stromer-Galley said she doesn’t want to ban its use in classrooms but instead be mindful about how students can use it appropriately.
“How do we encourage students to play with and use these technologies as part of an assignment and then critically think about it and reflect on it so there’s not a penalty for using it but an opportunity to think critically about what it means when you use it?” Stromer-Galley said.
Adam McCaffery — an SU senior studying public relations and information, management and technology — took Stromer-Galley’s IST343 class during his sophomore year. McCaffery recognized how artificial intelligence can be used to generate ideas for public relations campaigns but agreed that it needs to be used in moderation.
“It has to be a balance. Students, teachers shouldn’t rely on it. But I think that using it as a balance will really enhance education. Rather than … just relying on AI,” McCaffery said.
Schuster said he has concerns about AI as someone working in college admissions and about its presence in academia as a whole.
“Students are increasingly turning to these tools to produce term papers and projects, which, if undetected, can contribute to grade inflation,” he wrote. “Consequently, students may potentially gain admission or merit scholarships based on artificially inflated grades they did not genuinely earn. It is imperative to address and study these challenges for future cohorts.”
While Stromer-Galley encourages the use of AI, she also understands that students can sometimes abuse its power, leading her to contemplate the future of grading under new technological developments.
“We have to think differently about, what does that assessment look like, what are we really evaluating the students on? If it’s evaluating them on the writing, the likelihood is high that some of the writing is not the student’s own so maybe that is not the evaluation criteria, maybe it needs to be something else,” Stromer-Galley said.
Disclaimer: Adam McCaffery previously worked for The Daily Orange. He is no longer affiliated with The D.O. and did not have any influence on the editorial content of this story.
