Educators increasingly integrate generative AI (gAI) tools into academic learning, yet challenges specific to asynchronous online courses remain underexplored. This study examines the unique challenges of AI integration in online learning environments by comparing students’ performance in an AI-integrated assignment between online and in-person sections of a Food Regulations course.

Students in both the in-person (n= 19) and online (n=25) sections of Food Regulations were assigned to select a food law related to a historical event or social circumstance, create a report draft using a gAI tool and then revise it through critical evaluation, fact-checking, and additional research, and reflect on the benefits and limitations of AI in academic learning. Reports were graded on seven criteria: content clarity and accuracy, essay structure, citation, topic choice, AI tools and prompt use, ethical standards, and language/formatting.

Grade analysis revealed no significant difference in average scores (92.7 in-person vs. 86.6 online). However, reasons for point deductions varied. In-person students lost points primarily due to procedural errors, such as omitting AI prompts, neglecting reflections, or inadequate elaboration. In contrast, online students often submitted unedited AI-generated drafts, failed to fact-check, and cited suspicious or non-existent sources.

The disparity may stem from differences in communication. In-person students benefited from frequent discussions, Q&A sessions, and in-class reminders about ethical AI use. Online students received the same guidance through Canvas announcements and optional live Zoom sessions, but participation was limited.