Primary tabs

When students conscientiously object to using AI

By Eliza.Compton , 19 November, 2026
A new challenge is emerging for universities that require students to engage with generative AI: learners who baulk at using it. Before making AI compulsory in assessment, educators should ask whether its use is a learning outcome or merely a tool
Article type
Article
Main text

When I was a staff member at Murdoch University, the assessment policy allowed for “conscientious objection to assessment”. I thought this clause was strange because other universities, in my experience, didn’t have anything similar.

I asked colleagues at this public comprehensive university with some 24,000 students where the clause had come from and whether any student had ever used it to request a different assignment. None had seen a student conscientiously object to an assessment, and the general consensus was that the clause existed because of a concern raised by a veterinary student more than 30 years ago.

I now see students on social media regularly objecting to being required to use artificial intelligence in assignments on ethical grounds – for example, environmental impact, corporate greed, stolen content, copyright concerns and unaccountable billionaires. Some point to widely publicised examples of unsafe AI outputs and ask whether universities should require engagement with systems they do not trust.

Their most interesting objections, however, are often educational rather than ethical.

In my conversations with students, a growing number worry that AI undermines their learning. One student expressed the concern bluntly: “I’m doing a PhD to learn, not to have AI do the thinking for me.” Many describe generative AI as a shortcut they would rather avoid because they want to develop skills themselves.

Universities have spent the past four years wondering how to stop students using AI as a shortcut in assessments, while also directing them to develop AI skills. The issue of students refusing to use AI seems to sit between the two, and presents a challenge for educators: is AI a tool or a goal? And an important question then follows: how should we respond when students conscientiously object to using AI?

Decide the role of AI in the assessment or task

The first step is to determine whether AI use is a learning outcome that must be assessed.

In some assignments, the answer is clearly “yes”. Those studying computing, data science and software development might legitimately be expected to learn how to use AI tools effectively. If the aim of an assessment is to demonstrate AI literacy, requiring students to use AI is entirely appropriate. Their refusing to use AI in such a context would be similar to refusing to use a car to take a driving test.

In many assessments, however, AI is not the thing being taught, and its use is not the intended learning. Instead, it is merely one possible tool for completing a task.

Consider a report, essay or presentation intended to assess students’ understanding of course content. If the learning outcome concerns disciplinary knowledge, critical thinking or communication skills, educators should ask what educational purpose does requiring AI use serve? What learning outcome would be lost if a student completed the task without it?

This question is critical because examples are emerging of students being required to use AI for tasks that could be done through other means. In one online discussion, a student reported being instructed to generate both text and images using AI for an assignment, despite preferring to write the text personally and take photographs. If the assessment is designed to evaluate communication rather than AI use, accommodating the student’s preference will not undermine learning outcomes. Thus, teaching and assessment must distinguish between learning outcomes and learning tools.

Universities generally allow students considerable autonomy in how they achieve learning outcomes. Students may prefer different note-taking systems, writing processes, research strategies or productivity tools. Students are rarely required to use one research database to search for literature in preference to another. Educators typically focus on whether students have achieved the intended learning outcomes rather than prescribing every step of the process. AI should also be treated this way for most unsupervised assessments.

This does not mean, however, that all conscientious objections should automatically be accommodated. Alternative arrangements should be considered only when students can demonstrate the same learning outcomes without using AI and when accommodation does not create unfair advantages or disadvantages. Nor should educators assume that every objection is reasonable simply because it is sincerely held.

Rather, educators could approach these situations by asking three practical questions:

  1. Is AI use itself a learning outcome?
  2. Can the student demonstrate the required learning outcomes without using AI?
  3. Would an alternative tool or method compromise assessment standards, fairness or authenticity?

If the answers are “no”, “yes” and “no”, respectively, do not mandate AI use for the task.

How essential is AI to the task?

The broader lesson is that educators should be cautious about making AI use compulsory simply because AI exists. Universities rightly encourage students to engage critically with emerging technologies. Part of that critical engagement might include questioning whether, when and why particular technologies should be used.

The question is not whether students should learn about AI. They should. The question is whether they should always have to use it. Before requiring students to engage with AI, educators should be able to explain what learning outcome would be lost if students completed the task without it. If no clear answer exists, making AI optional is the more defensible choice.

Guy Curtis is associate professor in the School of Psychological Science at the University of Western Australia.

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.

Standfirst
A new challenge is emerging for universities that require students to engage with generative AI: learners who baulk at using it. Before making AI compulsory in assessment, educators should ask whether its use is a learning outcome or merely a tool

comment