Best way to fight AI in education? Remove grades altogether.
If AI has killed the university essay are grades still a valid measure of learning? "Ungrading" offers a different approach to tackling AI in education.
I have taken a small break from writing this blog, primarily because teaching has taken over my working life for the past two months. I wanted to give a short reflection on the use of AI by students and approaches taken in academia.
In my course on Ethics in cybersecurity, we initially set out by discussing the use of AI. Students and I seemed to agree that there were suitable uses for it - simplifying information for personal understanding, proof reading of spelling / grammar checking. Pleasingly many students didn't see the point in using to complete their whole assignment, saying that using it this way defeated the purpose of doing the course in the first place.
Even with this discussion, I did receive AI generated tasks for grading. Some obviously so, and others which I was suspicious of, but ultimately were inconclusive. For those who obviously used it, it came across as if students were less focussed on thinking through the topics of the course - and more about getting something which sounded correct.
This made me reflect upon the question: why are students motivated to use AI at all?1
Kui Xie an educational psychologist at Michigan State University, notes that if students are motivated to master a skill, there is no reason to cheat. However, if their primary goal is to appear competent, do better than their peers or just get the grade, then they are liable to use any tool available. So, for students to rely less on external tools such as AI there needs to be a shift from goal orientated to more intrinsic motivation. The best way for educators to try to get students to not use AI, is to get them to not want to use AI.
In looking further into motivation, I came across this quote from a post last year on the fantastic blog by Emily Pitts Donahoe:
If students are turning to ChatGPT to do their work for them, it’s because they see little value in that work beyond the grade they receive on it. It’s because attaining a certain grade, rather than learning, has become the purpose of their education.
The drive to cheat or use shortcuts has to do with the a) what the task itself is asking them to do, but also b) with what the current education system rewards. Pitts Donahoe and others write extensively on “ungrading”, an approach to education which takes grades out of the picture. There are several compelling arguments for doing this, but first I want to look at some of the issues around grading, which I think AI is further exposing.
The issue with grading
Educators would like to think that grades are both rewards for doing well and motivators to achieve. However, there are findings which suggest that grading actually does the opposite. Instead of motivating students to perform better, it enhances students' motivation to avoid receiving bad grades (Schinske & Tanner, 2014).2 Students focus their attention on avoiding punishment, rather than trying to achieve or better themselves through learning. There are other issues with grading, as it can:
dampen existing intrinsic motivation,
make students focus on extrinsic motivation,
enhance fear of failure,
reduce interest and enjoyment in class work,
increase anxiety,
lower performance in follow up tasks,
make students want to avoid challenging tasks,
and heighten competitiveness (Schinske & Tanner, 2014).
Furthermore, students getting a low mark might withdraw from class work, rather than be motivated to improve themselves. The fear of getting a low mark, can also be a motivator to cheat. Rather than motivate a healthy approach to work, grades enhance anxiety and avoidance of difficult courses.3 From first year university onwards, students become more “grade-oriented and work-avoidant and less learning-oriented.”4 In this kind of environment, and with pressure to achieve, it is no wonder that some students will take the easiest path.
A Short History of Grading
Students taking shortcuts to get grades isn’t a new phenomena. The education system puts such a heavy emphasis on grading, that it is no wonder students have learned to pursue grades while minimising effort. A fascinating article by Schneider and Hutt provides a history of the A-F marking system, primarily in European and American contexts. They write that the drive for grades created through the mid-20th century, created an ‘entrepreneurial student’, who would game the system by dropping difficult courses, seeking particular teachers or taking ‘easy-A’ courses. Over time grades became less a measure of a student’s pedagogical success, but instead a kind of currency through which to pass through the education system. They write:
“If learning learning sometimes had to occur along the way, so be it; but otherwise, students would do the least amount of work possible in order to attain the token of highest value.” (Schneider and Hutt, 2014)
Arguments arose in the 1960’s and 70's that extrinsic motivation (i.e. being motivated by grades) failed to focus students on the process of learning, and instead created unhealthy levels of competition and anxiety. One report in 1968 by Leslie and Stallings wrote that “an emphasis on grades encourages cheating, restricts study to material likely to be on the test, and encourages students to conform … to the instructors views and opinions”.
Even with the documented failings of a grading system throughout the decades, it still persists with schools and schooling depending on it determining placements and “communicating something about a student’s learning both internally within a school and externally to parents and other interested parties (Schneider and Hutt, 2014)”.
In this grade focussed environment, AI isn’t the root cause of cheating or short-cutting. It is just the latest extension of a system that doesn’t always reward the learning process.
The philosophy of “ungrading”
So, is ungrading the solution? Ungrading (or anti-grading) is an approach which does not reward grades for assignments or courses. It attacks the fundamental premise that giving grades is the best way to measure academic merit and to motivate students. The argument goes that, as educators, we have trained students to be in a mindset that getting good grades is all that matters. The purpose of education is shifted toward whatever gets the best score, rather than rewarding the process of learning. This is also a system which creates the motivation for students to cheat and deceive - to aim for grades rather than learn.
An anecdotal trend I have found in teaching philosophy is that some students who seem to get the most out of the course, do not always achieve the top marks. But these are students who I see developing new ways of thinking and looking at the world - which is exactly what I want the course to do. However, this development is also difficult to quantify in a numbered or lettered grading system.
Ungrading evangelist Jesse Stommel argues that most assessment mechanisms in higher education simply “do not assess what we say we value most". The process of grading also goes against many of the guiding principles educators go by, for example Stommel writes that there aren’t any university mission statements saying we should:
Pit students and teachers against one another
Rank students competitively
Frustrate learning with approaches that discourage intrinsic motivation
Reinforce bias against marginalized students
Fail to trust students’ knowledge of their own learning
All of these things, he argues, is what grading gives us. On this view, the traditional grading system is one which turns education into an adversarial practise, one where educators become characters in their own RPG fighting an onslaught of students hellbent on cheating, deceiving and generally tearing the flesh off of the corpse of academic integrity. The adversarial stance taken by some against AI is a continuation of an educational arms race that pits educator against student - but is what education should be about?
![DEEVICIS SOCE
- CHET
ι πιουο
C-PblJ
SHEO ΤΗΑ DEEVICIS SOCE
- CHET
ι πιουο
C-PblJ
SHEO ΤΗΑ](https://substackcdn.com/image/fetch/w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd4ec1e8b-f1ee-4a27-9252-b829a1c731f8_1792x1024.jpeg)
AI enhanced distrust in education
AI has only enhanced an atmosphere of distrust between educators and students. One study has found that more than half of 6-12th grade teachers reported Gen AI has made them more distrustful of whether a student's work is actually their own. Interestingly, schools where AI is banned report higher levels of distrust believing a students work is their own, compared to schools which allow its use.
Educators are probably underreporting the occasions where AI is being used. Some due to the burden of proof being so high; out of some Blackstonian ideal that it is better to let 10 guilty students escape than to punish one innocent one. Others might worry about their reputation of accusing the wrong students, such as the professor in America who flunked his entire class after incorrectly labelling them all of using ChatGPT. There is also the fact that doing a forensic analysis of every student is onerous, and which consumes time which educators simple do not have.
Some say they can spot an AI written essay, I agree to an extent. However, AI can get grades that are good enough. Ok it might not get the top marks that a comprehensive and well sourced paper will - but many students don't want that. If they can get a C then that is enough. However, the increase in capabilities of AI, and sites which promise to "humanise" AI writing will soon make this nearly impossible. One site called HideMy.AI offers to change your AI into type (essay, news article, blog post) and what level (Undergrad to PhD). The site is "on a mission to make all AI content read like a human wrote it."
Argument against
I can see the counter argument to the ungrading movement. Some might say not giving grades is like giving out gold stars to everyone just for participation. However when there are some who are using AI to cheat the system just to get gold stars, then they become inherently arbitrary and no longer a true marker of achievement or of learning. They have little basis of credibility for measuring anything beyond one's ability to get them.
There is some evidence that grades can motivate high achieving students, but at the same time, the can de-motivate under performing students. One study compared Grade 6 students who received grades and those who did not.5 It found that “graded low-ability students received lower subsequent grades through Grades 7–9 and had lower odds to finish upper secondary education, compared to ungraded low-ability students”.
Philosopher John Rawls, posits the fairest society as one in which aims to make the worst off better off - (this is the difference principle). With this in mind, is a system which motivates the least advantaged in education, a more fair approach to aim toward? Should merit be focused more on the effort of students who engage with the learning process, than with those who focus only on working for a grade? It is true that effort is not always rewarded in the 'real-world', but education is a space for personal reflection and growth. If effort is not rewarded here, then where else can it be rewarded?
Is ungrading the way forward?
The more I think about it, the more I consider that AI is exposing a major hole in the education system: that grades focus students on attaining an outcome rather than on the process of learning. This is not a newly identified issue, educators have long worried about the effect it has on student's motivations. Learning only enough to get them a high grade, or implementing strategies which will get the best grade. While it isn't new, AI has made this more apparent.
I am not sure whether ungrading is the ultimate solution, but what it does illustrate is the importance of a shift in focus toward the motivation of students. This is important in the AI debate. as we should consider how to make students want to learn rather than want to get marks or grades.
Perhaps the panic many academics and educators feel is from the incongruity of working within a system which is no longer able to measure what it sets out to measure - namely how much a student has learnt. The gravest fear is that it creates a new environment where neither educator nor student can put much value in the grade given or received.
I recognise that motivation also has to do with how assessment tasks are conducted, and the need to be constantly thinking of new ways to develop tasks to make them interesting for students.
Schinske, J., Tanner, K., 2014. Teaching More by Grading Less (or Differently). CBE—Life Sciences Education 13, 159–166.. https://doi.org/10.1187/cbe.cbe-14-03-0054
Chamberlin, K., Yasué, M., & Chiang, I.-C. A. (2023). The impact of grades on student motivation. Active Learning in Higher Education, 24(2), 109-124. https://doi.org/10.1177/1469787418819728
Kowalski, P. (2007). Changes in Students’ Motivation to Learn during the First Year of College. Psychological Reports, 101(1), 79-89. https://doi.org/10.2466/pr0.101.1.79-89
Klapp, A. (2015). Does grading affect educational attainment? A longitudinal study. Assessment in Education: Principles, Policy & Practice, 22(3), 302–323. https://doi.org/10.1080/0969594X.2014.988121