For Martine Peters, academic integrity looks like a logo that says a student used AI to complete an assignment.
“If you want to work with integrity, you have to be transparent about your AI use,” said Peters, who is a professor in the education department at the University of Quebec in Outaouais.
Peters has developed three logos that students put on their assignments. One says students did not use AI. Another says students used AI to help them. The final one says the assignment was generated by AI.
Her students include a list — separate from a bibliography — that says how they used generative AI and includes any prompts they used.
Peters is a director of the Partnership on University Plagiarism Prevention. The project involves researchers from Canada, the US and the UK, and focuses on preventing — rather than punishing — plagiarism.
Generative AI tools like ChatGPT have changed how students and professors at post-secondary institutions think about academic integrity. Plagiarism is no longer only claiming someone else’s words as your own, says Peters. It now includes using AI-generated words or images without acknowledging AI was used.
The dizzying hype surrounding ChatGPT’s entrance into popular use two years ago has faded. It is now accepted by professors and students alike as part of post-secondary education. But experts say more guidance is needed about how it is used.
Academics’ concerns about emerging technologies are nothing new, says Chad Gaffield, the CEO of U15 Canada, an association of Canada’s 15 leading research universities. Professors worried when search engines like Google or spell-check functions were introduced, for example.
“What’s different about generative AI, I think, is the ease with which it appears that something new is being created,” Gaffield said. “That’s the aspect that has caught people’s attention. But of course, as we know, it’s being created out of what already existed.”
Earlier this month, U15 released a paper outlining principles for how post-secondary institutions should use generative AI. Integrity and respect are cruciaI, it says.
The paper encourages universities to work with various groups on their campuses to develop policies and guidelines about using generative AI. The paper will be updated as generative AI changes, Gaffield says.
‘So tempting’
People can learn to use generative AI properly, says Gaffield.
AI only uses available information. If generative AI produces something that is inaccurate or biased against certain groups, it means it was working with inaccurate or biased information.
“Like all tools, like all technology, how [AI is] used is really the most important thing,” he said. “We have to … come at it in a way that does justice to what we’re trying to achieve in our educational institutions.”
Not many universities have official policies about proper AI use, says Phoebe Kang, an assistant professor at the University of Toronto Mississauga. Kang co-authored a study that examined the information that several Canadian universities have produced on generative AI. Some have guidelines, but guidelines are less authoritative than official university policies, she says.
“Policy comes with consequences,” Kang said. Universities may be hesitant to write official policies because they want more research, or because generative AI changes too rapidly for policies to remain current.
“A lot of responsibility lies with individual faculty members in how they make the decision on what is acceptable, what is not acceptable,” she said.
Professors need training about generative AI, says Faith Marcel, who co-authored the study with Kang and coordinates Niagara College’s Teaching English as a Second Language graduate program. “Continued professional development is really important to be able to get the word out there.”
“There are still students who are aware of AI, but are not exactly sure if they can use it or not. Faculty members are in the same boat.”
The U15 paper says clear communication is key. Students need to know what the expectations are in each course.
In Peters’ classes, students can use AI in a way that does not encourage plagiarism. She tells her students to ask AI programs to make lists or tables — but never text.
“Once you have a text in front of you, it’s so, so tempting to copy and paste it in your paper,” she said.
Students can also use AI to help with difficult tasks, like writing outlines, she says. “Using AI when it’s appropriate is a sign of intelligence,” she said. It shows students know which tools can help them improve their work.
Many students and professors share similar concerns about AI, says Mariel Miller, an assistant professor of educational psychology and leadership studies at the University of Victoria. Miller has surveyed undergraduate students about how they use generative AI for assignments.
Students had concerns about AI compromising academic integrity or providing inaccurate or biased information. They also worry about becoming too dependent on it.
Miller says AI can help review assignments, similar to peers giving advice. But while AI’s use will evolve, she thinks it should not replace experts — and that it cannot replace human interaction.
“Human interaction is enduring, and that’s always going to be a huge part of learning: being able to connect with people and people with different backgrounds and different experiences,” she said.
“I don’t think generative AI is going to replace that. I don’t think people want it to.”
