Updated: Sep. 9, 2025, at 3:13 p.m.
As artificial intelligence usage among students has boomed over the last two years, professors have utilized GW’s flexible AI policy to mold the technology’s role in the classroom how they see fit.
Given GW last updated its guidance on generative AI use in 2023 — during the technology’s earliest stages — faculty said they’ve crafted their own policies to adapt to how the technology has advanced in the past two years and its application to different academic fields has evolved. While some faculty have emphasized presentations and group projects to crack down on usage, others said they are encouraging their students to use AI as a tool for writing, studying and exploring their interests.
GW’s official guidelines for generative AI use, created in April 2023 by then-provost Chris Bracey, said it is permissible for students to use AI as a means to find ideas, but students are not allowed to submit AI-generated content for evaluation by professors.
In August 2023, the Office of the Provost released additional guidelines for faculty, encouraging professors to reinforce what plagiarism is, include a declaration of academic honesty on assignments for students to fill out and have students submit draft submissions so professors better understand what a student’s voice sounds like. The statement also affirmed that the guidelines are suggestions for professors to implement in their classes and are not binding policies.
University spokesperson Skyler Sales said individual programs may provide additional guidelines more tailored to how their department sees fit, while adhering to GW’s academic integrity policy.
“We will continue to update this guidance as needed and as technology, education, and research evolve, and we welcome input from the entire GW community as we work together to navigate and prepare students for our rapidly changing AI-enabled world,” Sales said.
Nicole Bartels, a professor of political science, said AI is on the minds of every professor on campus. She said after political science faculty noticed an uptick of students using AI on tests and written assignments in the spring semester, many within her department started holding conversations over the summer to discuss how to combat cheating.
She said ideas like verbally discouraging students from using AI and emphasizing in-class activities were among the ideas brought up in the meeting and that she has begun utilizing them in her classes.
“At the end of the spring semester, we all sort of indicated to each other that we needed to have a conversation about how are we going to deal with this moving forward,” Bartels said.
She said when AI first “made its entrance,” it was easier to distinguish between generated responses and a student’s work, but with newer versions of the technology it’s become harder for professors to detect.
OpenAI, the company behind ChatGPT, has continuously enhanced the chatbot since its 2022 debut. The latest version, GPT‑5, launched in August 2025, introduces major upgrades, including significantly improved writing capabilities and an optional “thinking” mode that allows users to trade speed for deeper, more accurate responses.
“AI gets better, it learns,” Bartels said. “So I think from two years ago versus where it is now, is a lot more difficult to make that very clear distinction of, like, this is AI.”
In 2023, professors’ AI said they were focusing their AI policies on redesigning their syllabi to prevent cheating amidst increased academic integrity violations. A 2023 study by Tyton Partners found 50 percent of professors were worried AI would negatively impact learning.
Bartels said she redesigned her syllabus this semester, requiring students to use a lockdown browser when taking assessments outside of class and implementing an activity where students are given an AI’s response to a question and have to mark up what the response got wrong or is potentially missing.
“I had to think outside the box of how better to assess students and what their knowledge was and not just what they fed into,” Bartels said.
Graham Smith, an adjunct professor teaching audio production, said he prohibits the use of AI in his class because he wants students to understand the basics of audio production before they start using AI to automate processes for them.
“I want people to come out of the class with real knowledge about how this stuff works, both from a technical point of view and from an aesthetic and psychological point of view,” Smith said.
He also said that this semester he began informing students of AI tools, like Descript, an editing program with features like AI voices and summaries, while still prohibiting them in his class, as he thinks they have become accessible enough for students to use if they want to continue with audio production beyond his class.
“It may have started to be good enough previously but not sort of accessible enough to where my students would have been able to easily tap into it and put it into their workflow, potentially,” Smith said.
But other faculty members encourage AI use more in their classrooms, including Tara Sinclair, the chair of the economics department, who said she views AI as a tutor students can ask questions to when they need help completing assignments for her courses.
Still, Sinclair said she prohibits students from using it for writing assignments, as figuring out how to convey thoughts is necessary to build critical thinking skills. She said she has moved away from longer, out-of-class writing assignments in favor of group projects with in-class components, which give her the chance to ensure AI is not used.
“To be able to learn how to think through things, you do have to go through the wrestling process of writing rather than skipping a step and having someone else do it for you,” Sinclair said.
She said as an economics professor, she fears many students taking economics courses for a general education or international affairs requirement may be tempted to rely on AI for their assignments since the coursework is more demanding and they might not immediately see the benefits the courses might provide. She said professors may need to adapt by becoming more “compelling” and work to convince students that classes are worth their time and effort.
“It is very tempting as a student to be like, ‘I don’t care about this class,’ ‘I haven’t been sold on the value proposition of this,’” Sinclair said.
Victor Krauthamer, an adjunct professor in biomedical engineering, said when students in his disability and rehab engineering course are working on their need statements — mandatory papers that identify a problem and propose a solution — students often limit themselves to solutions they are familiar with and generative AI can help identify a wider range of possible solutions to a problem.
Krauthamer said the University’s AI policy is “fine,” and he likes that the policy gives professors flexibility to teach how they want to, allowing faculty to integrate AI into their courses to a level they are comfortable with.
“There’s enough room for professors to teach the way they want to teach, which I think is very important,” Krauthamer said.
Rachael Jonassen, a professor for GW’s sustainable urban planning program, said she has a “love” for students feeding her assignments into AI programs to generate answers, as long as the final assignment they submit is creative, successfully synthesizes course material and cites credible sources.
“What I’m looking at is, how is the student expressing the synthesis or creativity that I’m asking for, if it’s well expressed and if it’s well documented with reliable sources, then it’s achieved what I think we need to try to be able to do,” Jonassen said.
This post was updated to correct the following:
The Hatchet incorrectly reported Sinclair said she feared international affairs students would use AI to cheat in her class because they were not interested in the economics department. Sinclair said she feared general education and international affairs students would be tempted to rely on AI due to the challenging course material and not recognize the benefits of the courses. We regret this error.
