”So… how was it?” “What did they ask?” In the past week, the sophomore section of the Stu was overflowing with one conversation: Milton’s new US History oral defense assessment. After a month of stressful research that culminated in a final paper, US History teachers announced this new addition to the assignment, which came as a surprise to students.
The format is straightforward: an individual has seven minutes to walk their teacher through their thesis, argument, sources, and research process. A student is expected to highlight any major discoveries or moments where their research direction took a turn. A brief three-minute Q&A session follows. As History Department Chair Joshua Emmott stated, these three minutes don’t aim to trip up the student: “They are questions to get the student to elaborate in more depth… They are not designed to point out what the student doesn’t know.” What Emmott did not mention in any of his answers was artificial intelligence.
Last year, according to The Milton Paper’s “State of the Acad,” a record-breaking 42% of students admitted to committing an AI-based academic integrity violation. A year before, though, in 2024, the percentage of admitted AI violators was 29%. With the rate of AI use rising, it’s clear that the timing of the oral defense rollout is not a coincidence. Although the History Department may not advertise its goals outright, it clearly crafted the oral defense to combat student AI use. The implication rings loudly: if you can speak for seven minutes about your paper, then you didn’t use AI. However, the oral defense, as it stands, might not actually accomplish that goal.
Emmott described the goal of an oral defense as giving students a chance to “come in and be able to be the expert and to talk about their research.” Nonetheless, this knowledge is exactly what someone using AI can fake. An oral defense, by design, asks students to re-read their paper, review their sources, and articulate their argument out loud. With a couple extra hours—and perhaps even less—of preparation, that rubric essentially provides a checklist that any student who used AI could complete. If the writer memorizes an AI-generated thesis, skims the sources, and practices speaking out loud, the format of the defense offers very little that could expose them. Meanwhile, the Q&A aims to check for gaps instead of inviting elaboration, meaning a student with surface-level knowledge of their topic can navigate the session comfortably. Students who used AI to write their paper after doing genuine research of their own would have even less trouble. Another aspect to be considered is the disadvantage of the Oral Exam for students who aren’t as comfortable speaking. If a student rushes, looks particularly nervous, or blanks while presenting, their hesitancy could be mistaken as a kind of confession to AI usage.
Although there isn’t a perfect solution to deter the use of AI, a pop-quiz style system could prove more effective. At an unannounced point in the research process, teachers could give a short quiz on each student’s topic. The absence of preparation is the point—if a student were to lean on AI, this method would expose them immediately. There are disadvantages, though; a pop-quiz would disadvantage students with test anxiety or those simply having an off-kilter day. Oral defenses posed the exact same disadvantages, but the unannounced quizzes serve as a more effective deterrent to AI usage. Perhaps combining the two would make for the best solution: keep the oral defense, but add an unannounced quiz somewhere to fill the gap.
The oral defense isn’t without merit. It develops skills that are valuable, like articulating your argument and thinking on your feet. However, if the History Department’s goal is to address AI usage, the oral defense alone won’t get them there.