At the University of Oregon every professor sets their own AI rules: students want a standard

By Zendoric · September 25, 2026.

As the academic year got under way at the University of Oregon (UO) in Eugene, local television station KVAL took the pulse of students on artificial intelligence in their studies: watchfulness, not rejection. The university does not impose a single policy. Each professor must state in their syllabus how they permit or prohibit the use of AI in their course, and is responsible for shielding their own exams against misuse of these tools. An 'AI Council', an internal body that advises the institution on the integration of AI across campus, completes the scaffolding, as KVAL describes it.

The voices gathered sketch a map of positions by discipline. Tiana Atkins, a fourth-year architecture student, calls for 'more regulation' and warns that outsourcing the creativity of the design process produces 'less human-oriented' results in a field that lives on individual judgment. Piper Such, a journalism student, admits to fear in the face of uncertainty: she does not know what shape the technology will take or how it will shape her career. Ayla Ballew, a first-year student, offers the opposite nuance: she used it in high school as a self-teaching tool, not as a substitute for her own work. Other students, according to the source, do say it helps them improve in subjects such as math and science.

As industry context, this decentralized model — leaving AI policy in the hands of each instructor, with an advisory council but no single rule — is the one many US universities are adopting in the face of a technology that changes faster than an academic regulation takes to be approved. It has the advantage of flexibility by discipline; it carries the risk of inconsistency: the same campus can treat AI as cheating in one class and as a recommended tool in the one next door.

Our reading is that the pattern emerging in Oregon reproduces, at classroom scale, something we have already seen sector by sector in the labor market: where the value lies in repeatable technical execution (solving an equation, debugging a calculation), AI is adopted without much drama; where the value lies in judgment and sensibility that are hard to replicate — architectural design, reporting with a voice of one's own — it generates more resistance and more demand for explicit limits. It is not a paradox: it is the same boundary that separates what can be automated from what still requires a human being behind it.

The job anxiety Piper Such expresses — 'we worry about our professional future' — is real and should not be minimized: this generation is entering the market just as the entry-level tasks that once served to learn the trade are beginning to be automated. That is the short-term problem, and it should not be built on false certainties. But it is worth keeping the wider horizon in view: if AI ends up freeing time and lowering the cost of routine tasks, judgment, distinctive creativity and human relationships — exactly what Atkins defends in architecture — gain relative weight rather than losing it. The university that today improvises rules case by case may be, without knowing it, already training the generation that will know how to demand that technology work for human judgment and not the other way around.