Students Embrace AI but Fear False Accusations

As institutions navigate the changing role of artificial intelligence in the classroom, many have implemented broad policy changes—ranging from restricting AI use to fully integrating it into coursework—with limited input from students. However, new survey data suggests AI is reshaping how students learn and work in ways that are not always reflected in the decisions colleges make.

In January, Packback, an educational technology platform that uses AI, surveyed nearly 700 college students across a range of institutions, including four-year universities, community colleges and online programs, about their use and views of AI. Respondents had all been enrolled in at least one Packback-supported course.

The report found that roughly 5 percent of students said they always or often use AI to generate a full assignment, a rate comparable to pre-AI forms of contract cheating, when students had to copy material from texts or peers. Of those, about 46 percent cited time constraints as their reason for using an AI shortcut, while 43 percent said they didn’t know enough about the topic and 42 percent said they lacked interest.

Kelsey Behringer, chief executive officer of Packback, said some students have always found ways to be academically dishonest, and AI is simply a new avenue.

“Students are using gen AI the same way a lot of knowledge workers do—as a thought partner or a brainstorm partner,” Behringer said. “I wouldn’t be surprised if some students were downplaying, so maybe that 5 percent is a little bit higher, but I think it’s consistent with the fact that the students who were academically dishonest 10 years ago are the same students who are academically dishonest today.”

The report found that nearly three-quarters of students are at least moderately concerned about being wrongly accused of using AI, including more than 40 percent who described it as a major concern. At the same time, about a quarter of students said they often use AI to support coursework—such as summarizing books, articles or other materials—while roughly three times as many believe their peers are doing so.

Behringer said fear of being unfairly accused is valid because instructors’ syllabi often include strict, zero-tolerance AI policies and rely on “flawed technology” to track usage, creating anxiety among students.

“What’s interesting is how little students admit to fully cheating with AI, versus how much they think their peers are fully cheating with AI,” Behringer said. “They’re also on social media, so they’re probably getting fed the scariest stories about AI infractions.”

“The risk is that we are deflating our students, demotivating them from pursuing education,” she added. “And I don’t mean to say this is all the faculty or administrators’ fault. I have a lot of empathy for their position—they’re trying to maintain academic integrity while balancing different stakeholders and emotions.”

Which students use AI: About two-thirds of respondents said AI helps them grasp complex topics and deepen their understanding, while more than half said it aids them in overcoming academic obstacles.

Students in business and management use AI more frequently across a range of academic tasks and are more likely to report significant benefits, according to the survey. That pattern may be partly driven by faculty in those fields, who are more likely to embrace AI and encourage its use than their peers in other subject areas, the report said.

By contrast, students in the social sciences and humanities said they were less likely to use AI often and reported fewer benefits.

Behringer said that while students often have to navigate inconsistent AI policies across courses, a one-size-fits-all approach isn’t realistic across disciplines.

“If we’re preparing [students] for what comes after college, whether that’s further education or a career, you’re not going to have consistent expectations across your organization,” Behringer said.

“I actually think there’s something to be said for how varying expectations in different courses prepare students for adaptability, understanding rules and context,” she added. “But the onus on faculty and the institution is to ensure there’s some sort of through line—a shared language across courses.”

Students’ role in AI policy: Behringer said the range of student reactions mirrors the range of reactions among faculty, underscoring the need for dialogue and collaboration in shaping AI policies.

“I cannot tell you how many students said [in the survey], ‘Burn it to the ground,’” Behringer said. “Maybe they have a passion for writing, or for knowledge, creativity or the novelty of thought, and they feel AI is offensive.”

“A lot of students want to reject it, but then realize, ‘Wow, I need to pay attention to this. I need to be [AI] literate because I’m going to be expected to, and I have to think about my future,’” she added. “I think that’s exactly how many faculty feel—like, ‘I don’t want to incorporate this into my work or my students’ workflows, but if I don’t, am I cheating myself and my students out of being future- and career-ready?’”

Behringer emphasized the importance of including students in these discussions.

“What you can learn from [the survey findings] is that students have to be a voice in the AI conversation, or you risk losing the people who are ultimately the heart and soul of the institution,” she said.

Get more content like this directly to your inbox. Subscribe here.