AI policy coming to University of Alaska as student misuse increases

As the use of artificial intelligence to cheat is becoming increasingly common among students at the University of Alaska Anchorage, the UA Board of Regents will consider adoption of a document guiding AI use across the university system later this year.

The UA system does not have a single policy governing AI use across more than a dozen individual campuses. A section in the student handbook prohibits "utilizing devices not authorized by faculty members."

The current guidance allows professors to tailor their syllabi to allow as much or as little AI usage in class as they like, but doesn't address broader concerns about how and where the use is appropriate on campus. The UAA Faculty Senate specifically requested the current structure.

"Allowing faculty to determine appropriate AI use in their classrooms was preferable to a single blanket policy because it preserved the integrity of academic freedom, allowed faculty flexibility, and enabled them to remain subject-matter experts in their disciplines," UAA spokesperson Catalina Myers said in an email Thursday.

Student use of the tools to cheat is a "steeply growing trend," according to a report on generative AI use among UAA students produced — with assistance from generative AI tools — by the Dean of Students Office and released in August.

"The data shows that there is a growing trend of reliance on generative AI tools in student work and a continued challenge posed in maintaining academic integrity in the classroom," the report said. "Cases involving AI use have taken over the majority of cases addressed by the conduct office."

A growing number of U.S. universities have adopted AI policies that are being "dramatically outpaced" by use of the tools, according to Stephanie Erickson, a special projects lead who presented the draft language to regents in September alongside Chief Information Technology Officer Ben Shier.

"We need to be deliberate with how we want to engage with AI instead of letting it just happen to us," Erickson said.

The university's delayed approach to regulating use of rapidly developing large language model tools has generated excitement among some faculty members, but some students feel the policy should be more stringent.

Regents will consider guiding principles for integrating AI throughout the UA system in November.

Shier drafted the AI intent language over the summer at the request of regents. Feedback from staff and students was included, but limited, according to Shier, who said the document is meant as a broad overview.

"This, in my mind, is one of the most profound technology shifts that we have ever seen. It's a critical topic that has a profound impact on what the university engages in," Shier said. "This document is very, very important."

Shier said the document outlines the university's AI ambitions while recognizing that developing the judgment, knowledge and skills of students is still its core focus.

Erickson and Shier presented principles, priorities and expectations for AI use in the document, which calls for "appropriate AI in every discipline" and emphasizes that humans must remain "in the loop."

"If there's one of these that might keep me awake at night more than others, it might be this one," Erickson said. "What we really want to determine is what systems are present where AI cannot make decisions. So 'in the loop' means that AI could recommend but a human decides. AI cannot act alone."

UAA recently added undergraduate and graduate degree programs studying AI, and the AI@UAA initiative under the college of engineering seeks to grow AI in Alaska, leveraging expertise in research and workforce expansion.

Aaron Dotson, UAA vice chancellor of research, innovation and economic development, said higher education's responsibility is to teach students how to learn, not policing the tools they learn with.

"There is no instruction manual," Dotson said. "We learn from experimentation, we learn from others and this is a venue to do that."

But students challenged by their classes have turned to AI chatbots in growing quantities. The report on generative AI use for academic misconduct at UAA shows a rapid increase since the tools were introduced. Of the 150 students cited for academic misconduct, 119 used AI, accounting for 1.1% of the more than 10,500 students who attended UAA last semester.

The report shows a 73% increase in AI use for plagiarism last year and a 56% increase in academic misconduct overall.

The report also said UAA requires all faculty to address AI use in their syllabus, but discourages the use of AI checkers. Andrew Harnish, an associate professor in UAA's writing department, said that in practice, many faculty use AI checkers to see if their students have turned in work generated by a chatbot, but use of AI isn't immediately disqualifying.

Student pushback

Brigitta Taaga, a sophomore studying legal studies at UAA who hopes to one day work as a paralegal, said she is worried that AI chatbots could make her future career obsolete. Taaga was introduced to ChatGPT during a study session with high school classmates

"At first, I thought it was really good resource, and then as I dived deeper into knowing the whole behind the scenes of AI, I found that it was also affecting our environment and our economy in a way, and it slowly pulled me from being so amused by it to being a little bit resentful towards it," Taaga said.

Taaga serves on the the student government for UAA students. During a preliminary meeting where the AI policy draft was presented to students, Taaga said she pushed back against the language, claiming it didn't go far enough.

Taaga told the regent's representative that when choosing what university to attend, she and other West Anchorage High School graduates were wary of schools with AI policies they felt were too weak.

"Most of the younger generations are really against it," Taaga said.

Michael Feign, a non-traditional UAA student, also criticized the draft policy language. Feign describes himself as "super geeky" and has decades of experience working with computers. He said he avoids using AI, and feels the draft policy fails to address privacy concerns.

"What they present is a lot of overarching ideas and not really like nitty-gritty policy," Feign said. "People have to fundamentally understand that it's just a tool, and the tool is what you make of it."

Feign said he would prefer the university protect student data by building its own large language model hosted on local servers.

He said the draft policy looks at AI with "rose-colored glasses" and he feels that a slim majority of students at UAA share his opinions. Feign suggested the university should add a prerequisite course on AI literacy to address the pitfalls of overuse before students become too reliant.