Initially, the introduction of ChatGPT alarmed many faculty members. The ability of an algorithm to generate a convincing essay in seconds raised concerns about the future of higher education and philosophy, which thrive on thoughtful reflection and well-crafted writing. Over the last two years, however, I’ve reached a different perspective: the issue isn’t that AI can write better than many students, but rather that it pushes us to define learning’s true purpose and how to impart it to students.<\/p>\n
In my philosophy courses at ASU, I’ve created several AI-enhanced assignments to address this. These tasks don’t aim to detect or prevent AI usage. Instead, they require students to employ AI in ways that render cheating almost pointless while ensuring learning. Each assignment positions AI as a partner rather than a threat \u2014 an endlessly patient interlocutor that absorbs reputational risks, promotes explanation, and encourages deeper understanding.<\/p>\n
The strategy is based on two key insights. First, as David Duran’s research review indicates, students learn more thoroughly when they teach by explaining, questioning, and clarifying for others. Second, Andrew Vonasch and colleagues highlight that people are so conscious of their reputation that they might endure discomfort to avoid appearing immoral. These insights reveal that students often remain silent not due to apathy but because they fear looking bad and facing social costs. AI, devoid of judgment and memory, offers a safer platform for students to express and share ideas.<\/p>\n
Philosophy involves rigorous inquiry, which requires vulnerability and a willingness to be wrong publicly. Students risk social capital when they face embarrassment, impacting their ability to cooperate and belong. In a reputation-driven classroom, silence is often the safest option. Conversely, large language models, lacking ego, can absorb the costs of trial and error, acting as a buffer and encouraging intellectual risks.<\/p>\n
Here are key assignments using AI as a learning catalyst, incorporating reputational psychology and the learning-by-teaching principle. In “Clap Back,” students debate an AI trained to defend specific philosophical positions like skepticism or theism. AI argues consistently, while students expose its reasoning flaws and defend a counter-position. This lowers reputational risks, enhancing philosophical skills, and increasing student participation.<\/p>\n
“Teaching for Botmerica” is based on Duran’s findings: students learn best when teaching. Students instruct an “ignorant” LLM on a philosophical concept, documenting their interaction for review. This assignment encourages metacognition, as students must understand material well enough to teach the AI, consolidating their knowledge.<\/p>\n
In “Reverse Office Hours,” students lead discussions while the professor questions them. The students’ performance influences the next lecture, encouraging preparation and making reputation a motivator for excellence. This captures the essence of learning-by-teaching, as students organize material and explain ideas to others.<\/p>\n
The “Rock \u2019Em, Sock \u2019Em LLMs” assignment is the most complex. Student groups train LLMs to defend opposing philosophical stances and participate in public debates. This exercise mirrors Duran\u2019s findings, as teaching demands reflective self-assessment, engaging students’ reputational psychology. This dynamic illustrates AI’s potential as a pedagogical tool for philosophical inquiry.<\/p>\n