Was the push to allow students to use AI in assessment a misguided response to the rapid rise of tools such as large-scale language models?
Here are four common arguments in favor of requiring students to be allowed to use AI, and why they're not as convincing as they seem.
“AI approval proves academic integrity”
One of the values of university guidelines for recognizing AI is that measurement of use supports: academic integrity. It seems easy to ask students to add an acknowledgment list. The student adds another table to the cover sheet, and the educator adds this as an additional criterion to the grading rubric.
Many universities allow the use of AI in insecure assessment tasks if students approve of it, but at the same time there is a risk of failure and academic misconduct if its use is not properly disclosed. Compounding this further is a mix of inconsistent messaging. Unreliable detection tools And a simple breakdown in logic means that this approach leads to sinister and counterproductive results.
One possible outcome is to allow students to create overall assessments and use AI. Another possible outcome is that the student could simply say, “I don't use AI,” and whether or not this is true, the student would still be meeting the requirements for approval. Although neither demonstrates academic integrity, both fulfill the requirements of the assignment through this acknowledgment.
Students are often reluctant to disclose their use of AI for the following reasons: vague guidelinesinconsistent enforcement and academic impact. This problem is made worse when teachers think they should mark students. more strict If you report (or suspect) that we are using AI. In effect, it punishes honesty, rewards deception, and risks creating distrust between students and teachers.
The bottom line is that AI acknowledgments cannot protect integrity. In fact, many universities are addressing integrity through secure in-person assessments that prohibit the use of AI, and public assessments that allow the use of AI. If these approaches already address academic integrity, acknowledgment simply adds a layer of procedure under the guise of responsible use of AI.
“AI approval shows students are using AI responsibly.”
In the first point above, our example was a student who admitted to abusing AI tools. The question arises: how can we know that students are using AI responsibly? After all, generating an entire rating is clearly not responsible use.
Acknowledging the use of AI creates the impression that students are using it responsibly. In reality, tables of AI prompts or long appendices of records are of little use in demonstrating judgment, evaluation, or professional standards when using AI. Universities cannot expect students to understand this intuitively. Universities have to teach it.
If acknowledgment is disconnected from the purpose of the assignment and learning, students will (understandably) view it as an administrative exercise rather than evidence of responsible use.
As for Demonstrate responsible use of AIjust as a checkbox declares that a student is the sole original author of a work, including a checkbox that says “I used AI in an ethical manner in the construction of this assessment” can achieve the goal of using AI in a principled way. Now, you might say this is an ineffective way to ensure responsible use of AI, and you'd be right. Both types of authorization are equally ineffective in ensuring responsible use of AI.
“AI endorsement teaches students how to use AI.”
Related to the responsible use of AI is the effective use of AI. The argument goes like this: By completing the acknowledgment list, students implicitly learn how to use generative AI in an effective way. They learn by doing.
Fundamentally, this is a misunderstanding of effective teaching practice. AI literacy If students are not clearly taught how to use the tools, there will be a lack of consistency. If students need to figure it out for themselves, we need to ask ourselves: How will this learning be evaluated? What are the standards for excellence in the use of AI? Will students receive meaningful feedback on their AI literacy and be able to improve? How much will acknowledgment affect grades? So what happens if students don't use AI at all?
You can't expect students to figure out AI on their own by looking at a list of uses for AI.
Instead, we need a programmatic approach to teaching AI literacy, which necessarily includes responsible use of AI. Students should engage in learning activities such as: critical thinkingEthical Reasoning and Evaluative Judgment. We need to help students not only understand how AI works, but how to use it responsibly in their academic and professional practices.
“The AI endorsement will allow us to see how our students are using AI.”
A common reason for creating an acknowledgment list is to help educators track student engagement. I have a lot of questions about student use cases. What platforms are they using? What types of prompts do they write? How often do they engage with it? This will be of great benefit to educators in guiding their work for the upcoming semester and adjusting teaching and learning to the known needs of their students.
There are advantages to using this acknowledgment. Having this baseline knowledge makes it easier to design for AI applications.
However, some issues remain. The first is that Alexandre Auguste Ledru-Lorrain, the leader of the French Revolution in 1848, said, “Go, my people. You must find out where they are going so that I may lead them.” We can now teach students how to use AI. If you wait until students are taught how to use AI, you risk missing the opportunity to co-design activities and create a shared understanding of the tool.
The second problem is that the data is of limited value. High-achieving students write pages and pages of usage that few in a particular group would imitate. The average student will histrionically (perhaps shyly) acknowledge commonplace usage, but that doesn't help the instructor's instruction. Unmotivated students may be better off just lying and saying they've never used it. We would have created a pile of work for everyone with little to show for it.
Creating teaching and learning activities that develop responsible and effective use of AI will give you a good idea of how your students will behave. Furthermore, assuming that students act thoughtfully and principledly, acknowledgments become unnecessary. Assuming that the student does not have the principles, the acknowledgment becomes just a false reference.
The transference of generative AI to education is frightening. We all want AI to be used successfully or to be phased out entirely. But at the end of the day, a piece of paper with acknowledgments on it alone doesn't ensure students can use AI effectively or ethically.
That job is on us.
Chloe Salisbury and Luke Zafir are Principal Learning Designers at the University of Queensland, Australia.
Want advice and insights from academics and university staff delivered straight to your inbox every week? Sign up for campus newsletter.
