The middle books of Plato’s Republic are full of smoke, mirrors, and philosophers. His Socrates toys with the interlocutor and reader alike, forcing both into a state of confusion about everything from romantic love to the role of ruling authority. This land of confusion is seen In Book VI, where Socrates elaborates on his assertion about philosopher rulers made at the end of Book V. He describes the kind of character and personality traits that are necessary for good leadership. These qualities, Socrates argues, are found in philosophers. The discussion spans paragraphs, but like any good teacher, Socrates provides a summary of his assertions at the end of the discussion. He states:
Is there any way, then, in which you could blame a practice like this that a man could never adequately pursue if he were not by nature a rememberer, a good learner, magnificent, charming, and a friend and kinsman of truth, justice, courage, and moderation? (Plato, 2016, p. 487a)
Looking at this list of virtues attained through education, we might find connections to institutional learning outcomes. Universities want students to gain comprehensive knowledge and develop skills to become lifelong learners. These goals easily mirror Plato’s own advocacy for good rememberers and fast learners. Universities want to produce civically-minded, responsible citizens. To put it another way, they want students to become kinsmen of justice and courage. Plato’s educated leader, hypothetically, is found in the students universities graduate.
In reality, Adeimantus’ skepticism of philosophic education mirrors contemporary reactions to the discipline. Socrates’ interlocutor, and real-life brother, argues:
He sees that all of those who start out in philosophy- not those who take it up for being educated and then drop it- but those who linger in it for a longer time-most become quite odd, not to say completely vicious; while the ones who seem perfectly decent, do nevertheless suffer at least one consequence of the practice you are praising-they become useless to the cities. (Plato, 2016, p. 487c)
Real philosophers, Adeimantus argues, are weird, vicious, or useless. They are usually a combination of all three. Thus the confusion looms for the reader. The philosopher, for Socrates, is the best leader. However, to Adeimantus, they are worthless.
There is a lot that should be unpacked about drawing lines of similitude between Plato’s claims about the whole person and contemporary conversations about pedagogy and the goals of university education. As a means to streamline the discussion, perhaps an analysis of a famous image will suffice. If one has spent any time in higher education they are likely familiar with the allegory of the cave. It is a popular image that is used in everything from teaching retreats to religious services. Trapped individuals are forced to stare at a cave wall and the shadows produced by puppeteers, until one is released and allowed to stumble into life outside the cave (Plato, 2016, pp. 514b–517a).
Of relevance to this discussion is Plato’s use of puppeteers. Plato doesn’t specifically call these people “puppeteers:” Instead, he simply describes individuals who are carrying objects that produce shadows that the cave dwellers mistake as real. The result is a tragic life of the dwellers. They are completely content, but unaware of what is reality. They live, but they don’t live well. As Socrates says, “such men would hold that truth is nothing other than the shadows of artificial things” (2016, p. 515c). The puppeteers are never described as ‘bad’ or ‘evil,’ but rather just perpetuate shadow production. In fact, there’s no discussion on whether the puppeteers are aware of the world outside the cave either. Plato’s ambiguity in this description gives the reader an opportunity to fill in the gaps with their owned lived experiences. For example, one might impose into the brief allegory that the puppeteers are just as trapped as those who stare at the wall. This creates a sympathetic interpretation of the puppeteers, making them victims of the cave as well.
The cave image provides a framework for addressing current discord about generative AI in the classroom. The easy connection to make might suggest that these AI programs are examples of ‘artificial things manufacturing shadow,’ or that artifacts generated by these tools are shadows meant to represent student learning. Such a comparison seems flawed. Students have always been the trapped cave dwellers. They are the ones who are moving from a state of cave-based ignorance to glimpsing reality. Instead, it could instead be suggested that the ‘artificial thing’ is educational assessment. Our systems of student evaluation are shadow manufacturers, and educators are the puppeteers. To understand this criticism, there must be reflection on current assessment practices. What are the goals of course assessments? Why are papers and exams such a significant component of university humanities assessment?
A simplified response might be that these tools serve as a means to evaluate student learning. Papers, for example, exist so that we can observe students’ abilities to research, vet sources, develop arguments, and communicate clearly. The goals of these evaluations seem noble, but there are two glaring issues. First, these methods of assessment are inherently flawed in determining actual student learning. Exams are neither necessary nor sufficient in revealing what a student learned. Instead they show what a student was able to recount under specific circumstances. Students cram and then forget soon after the exam is over. Students suffer from test anxiety and ‘blank’ even though they’ve thought deeply about a topic. They endure traumatic episodes before sitting down to take the exam and cannot focus. Exams are an incomplete means to evaluate student learning and understanding. To return to the Platonic image, exams and similar assessment methods, are reflections and shadows of what a student understands. Such a conclusion isn’t particularly revolutionary, as there has been decades of research indicating the concerns of using exams as learning indicators. A report from the Economic Policy Institute describes the situation succinctly:
Nonetheless, there is broad agreement among statisticians, psychometricians, and economists that student test scores alone are not sufficiently reliable and valid indicators of teacher effectiveness to be used in high-stakes personnel decisions, even when the most sophisticated statistical applications such as value-added modeling are employed (Baker et al., 2010).
It is known that exams are insufficient, and that utilizing other means of assessment is necessary to fully articulate student learning.
What is concerning many humanist educators is that AI reveals that papers, by themselves, are not indicative of student learning either. Recent surveys have shown that students have a growing familiarity with these tools, and many are using them to complete assignments. A majority of students responding to the Best Colleges survey consider the use of such tools cheating (Lyss, 2023). Again, academic dishonesty isn’t a revolutionary idea. Paper writing services are prevalent and quite lucrative for graduate students looking to make some money. AI didn’t create student dishonesty. It made it free, readily accessible, and extremely difficult to catch. The ease of use and cost-free entry point made the possibility of ‘effortless papers that can’t be detected’ a reality for students regardless of economic level. And with this new opportunity of easy grades, it is obvious why students would succumb to the temptation. After all, it is the evaluated artifact that is pushed to matter, not the process undertaken to produce the artifact. Classroom evaluation methods are shadows. They provide glimpses and possibilities, but are nothing but incomplete sets. They show a product, not a process. And GPT has shown that created artifacts are neither a necessary nor sufficient indicator of student knowledge.
There is a second concern here that seems deeply systemic. This issue isn’t about what students produce, but rather why they produce it. Returning to the previous discussion on assessment, it is clear that artifacts like papers and exams exist to assess the development of student knowledge and learning. This evaluative process takes time, energy, and failure. But, when observers look at everything from student conversations to faculty evaluations, the focus is not on the act of learning but rather the grades that students receive. The artifact, the grade, is what matters, not the intellectual development which produced the grade. For all of a university’s discussion on student learning, it is not the learning that actually matters: it is the product, the grade. And in focusing on the product, important learning avenues are devalued or removed altogether. Failure, for example, can be a vital component of memory and learning. As noted by Roediger and Finn, “Pupils actually learn better if conditions are arranged so that they have to make errors. Specifically, people remember things better and longer if they are given tests so challenging that they are bound to fail.” (Roediger & Finn, 2010). When the concern is simply the letter grade, then the learning garnered from failure is not considered at all as part of the grade.
There are two things of note here. First, this ‘product focused’ mentality is perpetuated by the university system. Gaining access to student aid is tied to product. A student must receive a certain grade to be eligible for scholarships or government backed loans. Access to major classes is tied to product. Graduation honors are tied to product. Being retained as a faculty member is tied to product. And these products which are supposed to represent the process of learning are Platonic shadows. Tests, papers, projects, and other means of assessment are incomplete, vague effigies of student learning. For institutions who have learning outcomes based in the language of student skill acquisition, much of the evaluative process is tied to artifacts that do not necessarily represent that achievement. ‘Evaluative process’ remains intentionally vague. It applies to both students and instructors. Students need sufficient grades to graduate, and instructors need high student grades for strong tenure portfolios. The system drives students and faculty to focus on the product of a letter grade.
Educators must participate in this cycle, as grades and graduation rates are tied to everything from diplomas, to tenure, to regional accreditation. Many of our systems are supported by signs and not by actual student learning. The result is another cave and another land of shadows. To put the discussion another way, consider this assertion by Rick Wormeli, “We want an accurate portrayal of a student’s mastery, not something clouded by a useless format or a distorted snapshot that doesn’t represent true proficiency. The best phrase to apply here is ‘clear and consistent evidence’” (Wormeli, 2006, p. 31,32). However, the presence of AI has shown us that written artifacts are neither clearly students’ work nor are they consistent indicators of actual human learning. Notice Wormeli’s description of frivolous assessment. He calls it “clouded” and “distorted.” These assessments are shadows that no longer fail to address employers’ needs, but fail to represent actual student work.
Second, it is suggested that there is a place where prioritizing product is both expected and warranted. In the world of business, work is tied to product: what someone sells, what someone maintains, what is created. Streamlining production is often seen as a good in this sphere. “Work smarter, not harder” becomes the mantra. The product is the goal. Contrast the objectives of this sphere with that of a healthy personal relationship. Being a good significant other, parent, child, or friend is tied to the process: being present instead of producing. For Plato, at least, it was the process of undergoing improvement that was actually the goal of improvement. As Kenneth Henwood observed, “The individual in his moral life, like the artisan in his productive life, faces an endless task of betterment. No artisan, whether in the productive or in the moral life, achieves the discernible perfection of his art” (Henwood, 1979). The perpetual process of improvement that is vital for experts. Using the example of the art, the individual pieces produced pale in comparison to the drive to improve one’s knowledge and work. It was the act of techne that made learning experiences authentic. This conclusion is seen in some contemporary discussion on design education:
The traditional studio model of graphic design education approaches student learning through immersion in practice, imagining novice designers absorbing abilities and dispositions through experimentation and proximity to discipline experts and peers. If this model ever truly operated in Australia’s design schools, it has been eviscerated by the massification, marketization and rationalization of tertiary education since the early 1990s (Wragg et al., 2023).
While the conclusions drawn by Wragg about the massification of design education are beyond the scope of the current discussion, the emphasis on immersion, practice, and proximity to experts is not. Finding ways to pivot away from an emphasis on products affiliated with grades and instead focus on the practice of our disciplines should be the goal of humanist educators. Because of this, humanities teachers should look to incorporate opportunities for producing safe, reflective space to immerse students in the process of critical thinking. Such a conclusion is not revolutionary, but thinking about how AI can provide that safe, reflective space could be.
Online philosophy classes at Utah Tech are beginning to explore how AI can provide safe discussion space for students. Philosophy challenges students’ deeply held beliefs, and so it is at times difficult to create a truly open space where students can reflect on course material. Online discussion threads, for example, often create environments of self-censorship, as students are worried about seeming underprepared or foolish. AI provides the means to practice the dialectic in a place that removes other students’ judgment. Consider, for example, the following scenario and assignment. Imagine students are tasked with asking a text generator about the three definitions of justice provided in Book I of the Republic. These definitions are well-known, and the program is not likely to generate false information. From there, students might be encouraged to press the program about errors of the definitions, or even debate with the program as to what definition is best. The assignment would expect students to ask the program three to five questions, and then analyze the AI responses. They would also be asked to reflect on how the responses of the AI contrasted their own conclusions about the text. Students would then return to this assignment at the end of the semester to see how their understanding of that AI response has changed since encountering the whole Republic. Students will use this assignment to see their learning process, contrasting where they are at the end of term with their early semester observations.
AI presents unique learning opportunities. It provides accessibility and a safe space for students to ask questions without fear of social ramification. It also, however, shows us foundational, systemic concerns about why and how we, as institutions, assess students. Like Socrates, our goal should be becoming ‘battle partners’ with students. This partnership hinges on moving away from production-focused thinking that centers on shadow artifacts. AI, therefore, is far from the ‘boogeyman’ it initially appeared to be. It might actually end up being a tool that allows students to become stargazers.