As the influence of large language models (LLMs) like ChatGPT continues to grow, Princeton University is at the forefront of exploring the intersection of artificial intelligence and education. Some professors at the prestigious institution are harnessing AI as a pedagogical tool, while others are treading cautiously amid concerns about academic integrity.
“Blockie” – ChatGPT on steroids
Professor Pramod Viswanath, an Electrical and Computer Engineering professor at Princeton, has introduced his own AI teaching assistant, affectionately dubbed “Blockie.” Blockie takes inspiration from ChatGPT but is tailored for advanced engineering classes. Viswanath describes it as “ChatGPT on steroids.”
Blockie is trained on lectures and notes from Viswanath’s advanced engineering course, offering personalized assistance to students. While AI tools have faced varying degrees of acceptance among Princeton faculty, Blockie’s emergence underscores the potential of AI in education.
Princeton’s stance on AI tools
Princeton’s Office of the Dean of the College and the Office of the Dean of the Graduate School sent a memo in January 2023, titled “AI & ChatGPT Guidance for Teaching,” acknowledging the University’s openness to AI tools in the classroom. The memo emphasized that each faculty member has the freedom to use ChatGPT and similar technologies but must adhere to academic integrity rules and collaboration policies under the University’s Honor Code and Academic Regulations.
This memo reflects Princeton’s willingness to explore AI’s role in education while maintaining academic integrity.
ChatGPT: Limited but controversial
Despite the buzz surrounding ChatGPT, opinions about the technology vary among Princeton faculty. Some see it as limited, producing text that may initially appear convincing but ultimately lacks substance. Matt Weinberg, an assistant computer science professor, characterizes it as producing “text that initially seems convincing but is ultimately BS.”
Concerns about overreliance on AI tools have led to bans in certain classrooms. For instance, students are generally prohibited from using ChatGPT in Princeton’s introductory computer science course, COS126: An Interdisciplinary Approach. This course aims to teach students fundamental concepts independently, without AI assistance.
Sebastian Caldas, a new lecturer for COS126, highlights the importance of students learning essential concepts on their own. He believes that introducing AI tools too early might hinder their foundational understanding.
Blockie’s role in advanced courses
Blockie, the brainchild of Viswanath and his teaching assistant Tianle Cai, steps in to assist students in more advanced courses like ECE 470: Principles of Blockchains. This course delves into the intricate design and algorithmic principles of blockchains, a technology essential for secure data sharing.
Blockie was trained using lectures and assignments from ECE 470, making it a personalized teaching assistant for students in the course. While ChatGPT might excel in introductory courses, it struggles to provide relevant assistance in advanced classes, which is where Blockie fills the gap.
Student perspectives on blockie
Students taking courses with Blockie have recognized its value. It streamlines office hours by providing answers to common queries and clarifications on assignments. However, students emphasize that Blockie’s usefulness depends on their existing knowledge of the subject matter. It complements their understanding but doesn’t replace the need for a solid foundation.
Jonathan Jeong, a student in the course, finds Blockie particularly helpful for repetitive tasks and simplifying complex problems. Joy Patterson, another student, describes Blockie as “invaluable” for learning Rust, the novel programming language taught in ECE 470.
An evolving educational landscape
Princeton’s willingness to explore AI in education extends beyond computer science. In courses like POL386: Violent Politics, students are encouraged to use ChatGPT to generate responses and engage with AI as part of their learning experience. Professor Jacob N. Shapiro believes that exposing students to AI tools prepares them for the modern workforce, where such technologies are increasingly prevalent.
Shapiro trusts Princeton students to responsibly engage with AI and expects them to cite AI-generated content. He sees AI as an opportunity for students to hone their critical thinking skills and distinguish between coherent and nonsensical AI responses.
Looking ahead
As AI continues to evolve and find its place in education, Princeton University remains at the forefront of this transformation. The integration of AI teaching assistants like Blockie and ChatGPT represents a significant step toward exploring the possibilities of AI in the classroom.
While challenges related to academic integrity persist, Princeton’s approach encourages responsible AI use and fosters a learning environment that prepares students for a future where AI tools are commonplace.
As Professor Pramod Viswanath aptly puts it, “This is happening, there is no question about it. Every class has to have an AI assistant, I’m willing to go on record for that.”
As Princeton and other institutions adapt to this new educational landscape, AI’s role in shaping the future of learning is becoming increasingly clear.
