Alaska’s Draft AI Policy Says a Human Decides. Your Syllabus Should Say Where.

Alaska’s draft AI policy says a human decides. Your syllabus should name which human, at which step, before a student ever asks.

On October 4, the Anchorage Daily News reported that the University of Alaska Board of Regents will consider guiding principles for integrating AI throughout the UA system in November. Today the system has no single AI policy across its more than a dozen campuses. What it has is a student handbook line that prohibits “utilizing devices not authorized by faculty members,” and a great many individual syllabi.

Most coverage will file this AI policy story under governance. Read it as a faculty member and one sentence in the draft does more work than all the rest.

The sentence that defines human in the loop

The draft was presented to the regents in September by Stephanie Erickson, a special projects lead, and Ben Shier, the system’s chief information technology officer. Shier wrote it over the summer at the regents’ request. It calls for “appropriate AI in every discipline” and says humans must remain “in the loop.”

Plenty of policies say that. Erickson explained what it means, which is the rare part.

“What we really want to determine is what systems are present where AI cannot make decisions,” she said. “So ‘in the loop’ means that AI could recommend but a human decides. AI cannot act alone.”

She also said it was the principle most likely to keep her awake at night. It should keep the rest of us up too, because it is not a sentence about students. It is a sentence about every place in a course where software touches a decision about a student, and in most courses some of those places belong to the instructor.

Where the AI policy stops and your syllabus starts

Here is how Alaska works today. The current guidance lets professors tailor their syllabi to allow as much or as little AI use as they like. The UAA Faculty Senate asked for that structure. A university spokesperson said it “preserved the integrity of academic freedom” and “allowed faculty flexibility,” and that it was preferable to a single blanket policy.

That is a reasonable design, and it puts the weight on one document. According to a Dean of Students report on generative AI and academic misconduct, released in August, UAA requires all faculty to address AI use in their syllabus.

So whatever the regents adopt in November, the syllabus is still where a student finds out what the rule is in this course, on this assignment. A system principle can say a human decides. It cannot say that, in your course, the human is you, and that the decision happens before any grade changes.

Your syllabus probably covers half the loop

Most syllabus AI policy statements answer one question: what may students do with AI? That matters. It is also half of the loop.

The other half is the instructor’s side. An AI checker returns a score. A feedback tool drafts comments. A grading assistant suggests a mark. An early alert flags a student who stopped logging in. Every one of those is AI recommending. The question Alaska’s draft AI policy forces is simple. Who decides what happens next, and does the student know?

The UAA report already shows the gap. It discourages the use of AI checkers. Yet Andrew Harnish, an associate professor in UAA’s writing department, told the paper that in practice many faculty use them, and that use of AI is not immediately disqualifying. Read through the draft’s definition, that practice can be sound. A checker can recommend. A person decides, with more than the score in front of them. The problem is not the checker. The problem is that nobody wrote down which one is which.

There is a quieter example in the same story. The misconduct report itself was produced with assistance from generative AI tools. That is not a scandal. It is the definition working as intended, as long as a person decided what the report says. Your course deserves the same clarity about its own tools.

The numbers behind the urgency

The report gives the regents reasons to move. As the Anchorage Daily News describes it, of the 150 UAA students cited for academic misconduct, 119 used AI, which the paper puts at 1.1 percent of the more than 10,500 students who attended UAA last semester. We have not seen the report itself, so read those figures as the paper’s account of it. By the same account, the report also shows a 73 percent increase in AI use for plagiarism last year and a 56 percent increase in academic misconduct overall. In its own words, “Cases involving AI use have taken over the majority of cases addressed by the conduct office.”

Hold both numbers at once. AI cases dominate the conduct office, and the students cited are a small share of the student body. An AI policy written only for the first number treats every student like a suspect. An AI policy written only for the second misses what the conduct office sees every week. The recommend and decide line works for both, because it does not ask whether a student is guilty. It asks who decides, and on what evidence.

What students said about the draft

Students did not all want the same thing. Brigitta Taaga, a sophomore in legal studies who serves in UAA student government, pushed back at a preliminary meeting where the draft was presented, saying it did not go far enough. Michael Feign, a nontraditional student, said the draft fails to address privacy and offers “a lot of overarching ideas” in place of specific policy.

Both criticisms land on the same spot. Principles are the right level for a university system. Students live at the level of the assignment, where “a human decides” either names a person and a step or names nothing. Shier said feedback from staff and students was included in the draft, but limited. Your syllabus is where that feedback can still be answered, one course at a time.

What to write before the regents vote

You do not need to teach in Alaska to use this. If your institution’s AI policy has no definition yet, borrow this one. If it has one, check whether your syllabus says where it applies. Three moves, for one assignment.

One. List every place AI touches the assignment. Both sides. Where students might use it, from brainstorming to the final polish. Where you might use it, from a checker to a feedback draft to a reminder email. Most instructors find more on their own side than they expected.

Two. Mark each one recommend or decide. For every touch point, say whether AI may recommend there or is not used at all. Then write down who decides. If the honest answer is the software, that step needs a person.

Three. Write the clause. Two sentences for the syllabus. One tells students where they may use AI to get a recommendation and what they must decide themselves. One tells them where you use AI and that no grade, flag or finding happens until you decide. Students are owed both sentences before the first submission, not after the first conduct referral.

Aaron Dotson, a UAA vice chancellor, put the state of things plainly: “There is no instruction manual.” That is true for the system. For one course, you can write the page that matters.

An AI policy can tell your students a human decides. Only your syllabus can tell them which human, and when.

Your next step

If you want the fastest version of this, take one assignment through The Human Decides Clause, by hand, in about 25 minutes.

When you finish you will hold a map of every AI touch point in that assignment, the decisions only a person makes, and a two sentence clause ready for your syllabus. Then the next thought arrives on its own. That is one assignment. Your syllabus covers every one of them, in every section you teach.

That is the job the Syllabus Generator does. It builds a student centered, accessibility first syllabus grounded in your Course Record, the five minute record of your course, your outcomes, your policies and your voice that every tool in the toolkit uses. Put the clause in your Course Record once and it is there for every syllabus you build.

If you would rather think it through with other instructors first, the Harvest Kernel Learning Community is free, and this question is the conversation there this week.

Sources

• Tim Rockey, “University of Alaska regents weigh AI policy as student misuse increases,” Anchorage Daily News, October 4, 2026. adn.com