Byron Obed Sagastume Bran/iStock/Getty Images Plus
Until last semester, neither of us had a student ask about how to proceed if they didn’t want to do an assignment. That was until Carla designed an assignment requiring both use of and critical reflection on generative AI.
We are two philosophy professors who care about pedagogy but who have different opinions about generative AI. Jenny vehemently rejects all use of generative AI, while Carla uses it judiciously for some aspects of academic work. Carla had heard stories about students using generative AI “all the time,” whether or not it is allowed, and our university encourages faculty to include generative AI in our pedagogy and course outcomes. She thought that requiring students to use it and reflect on this use would provide them with firsthand knowledge of its costs and benefits.
Carla’s assignment, embedded in a course that included a focus on research methods, involved group projects to create a presentation and an essay about the life and work of a significant scholar. She required students to use generative AI, drawing upon guidance about effective usage. She also required that they engage with criticism of generative AI and submit a paper reflecting on their use of it.
When Carla explained the assignment, the first question raised was “What if we don’t want to use GenAI?” At first, she responded that it was part of the assignment and that whether or not they wanted to use generative AI, they should at least know how to use it. If they then chose to reject it, they could do so from a place of knowledge and experience.
The class atmosphere became tense as more students pushed back against this required use of generative AI. They wanted to know what their classmates thought rather than what a large language model produced in response to prompting; they wanted to wrestle with difficult readings and complex ideas; they wanted to work together across their human differences to negotiate new ideas and find ways of expressing them. And they didn’t want to outsource any of this effort to generative AI. They wanted to do the work, even the mundane work.
Because these concerns were deeply held by more than a couple of students, Carla amended the assignment and gave students the option to be conscientious objectors. If students had strongly held reasons for eschewing the use of generative AI, they were free to complete the essay and presentation using only traditional means. Instead of writing a paper reflecting on their use of generative AI, they could write a paper explaining why they chose not to use it and the costs and benefits of that choice.
Seven out of the eight groups took the conscientious objector option and decided not to use generative AI. The eighth group experimented with it but decided they preferred to work without it. In their essays, students expressed deep concerns about cognitive dependence, environmental impact, racial and gender bias, misinformation, academic integrity, and worker exploitation. Students were invested in their own learning, developing their creative identity and fostering genuine intellectual relationships with each other. Students read about and discussed generative AI and made carefully considered decisions not to use it.
We’ve been told that all students use generative AI now, that they don’t know how to read and write unassisted and that they don’t want to. But Carla’s students show us that this isn’t right. And they’re not the only ones. Jenny has had students, of their own accord, tell her how much they hate the expectation that they will, or should, use generative AI—for all the same reasons. One told the story of an in-class discussion of generative AI in which he was told by a professor that he must be lying when he said he never used it. We need to listen to what these students are saying, not dismiss them as liars or simply assume that they don’t exist.
And these students are not outliers. Young people are very concerned about generative AI degrading their cognitive and social capacities. A recent survey of young Americans under 30 found that 61 percent worry that the use of AI will make people worse at thinking creatively, and 58 percent worry that it will impair the ability to form meaningful relationships with others. Another survey shows that 31 percent of Gen Z respondents reported “anger” about AI, a very large increase from 22 percent the previous year.
Our students—and others in their AI-skeptical generation—are thoughtful moral agents who have used their capacity for critical reflection to arrive at compelling and deeply felt moral objections to using generative AI in their university work. We think educators should respect this. First, they need to be aware that students like these exist. And second, we think they should seriously consider honoring it—either by simply not requiring use of generative AI or by creating alternative assignments for students who declare themselves as conscientious objectors.
For some educators, their learning objectives actually require the use of generative AI by students, and making accommodations like this won’t be possible. For others, it may be prohibitively difficult. Our point is not that everyone must create room in their courses for conscientious objectors, but rather that everyone should be aware of their existence and give serious thought to honoring their opposition to generative AI.
What do we mean by saying that students are conscientious objectors? We looked at the Canadian Armed Forces definition of conscientious objection: “A conscientious objection is a sincerely held objection, on grounds of freedom of conscience or religion, to participation in: (a) war or other armed conflict; or (b) carrying and use of weapons as a requirement of service in the CAF.” Freedom of conscience is defined as “protecting views based on strongly held moral ideas of right and wrong.”
Our students who reject generative AI are very clearly doing so on grounds of freedom of conscience. Such rejection is not just available with respect to war, but also, for example with respect to health-care professionals in relation to medical assistance in dying (in Canada, MAiD). Our students are not merely raising some idiosyncratic concerns. As with MAiD and warfare, many extremely thoughtful people have good reasons for objecting. Indeed, experts across many fields have argued strenuously that the adoption of generative AI brings with it serious harms. And the legitimacy of conscientious objection has begun to be recognized in the workplace: For example, engineer Rachel Maus has obtained a religious exemption from using generative AI.
But, we are repeatedly told, employers want students with skill at using generative AI. This doesn’t mean, however, that this skill needs to be taught in every class or to every student. Employers want many things, and we are not all responsible for teaching all of them. Many employers want employees who can use a spreadsheet, yet we have never taught that in our philosophy classes. Students can, and indeed must, learn different skills in different classes. Moreover, the non-AI skills that we teach remain valuable. To take the skills we teach in our own field as an example, students need to be able to read closely and carefully, to communicate across differences, to genuinely understand other points of view, to recognize good objections to ideas, to critically reflect on what they are being told, and to persuasively argue for their own views. In fact, in an era of inevitable AI hallucinations and persuasive yet terrible AI advice, these skills are becoming more rather than less valuable.
And employers are increasingly recognizing the fallibilities of LLMs and the importance of human skills. One study found that 95 percent of businesses are getting zero return from their investments in AI. As it is integrated into workplaces, workers who can read and think carefully, assess accuracy, consider ethical issues, and discuss potential problems only become more important. These are precisely the skills that we teach through the slow processes of reading, thinking, talking and writing for oneself. The conscientious objectors stand to excel at these skills.
Importantly, we are not arguing that anyone should be required to recognize students’ conscientious objections to using generative AI. We strongly believe in academic freedom and maintain that educators must be free to set their own learning objectives and dictate the methods to be used to achieve them. What we are arguing is that we—both individual academics and academic institutions—should be aware that conscientious-objector students exist. We should stop assuming that all students embrace and use this technology. And we should all consider whether we can make a place for conscientious objectors in our classes that honors their concerns and their autonomy.
