A Penn professor used AI to replicate part of a master's course — and says it threatens universities' business model

In a groundbreaking exploration of artificial intelligence in education, Jesús Fernández-Villaverde, an economics professor at the University of Pennsylvania, has revealed the transformative potential of AI in replicating a master's course. Utilizing Anthropic's chatbot, Claude, he completed a personalized study on sociologist Erving Goffman in just 12 hours, a task that typically demands a week of intensive study. Fernández-Villaverde crafted a tailored syllabus with Claude's assistance, covering essential readings and key concepts akin to a traditional master's program. "What I achieved in 12 hours mirrors what students usually spend 9 to 12 hours on in a structured course," he stated, emphasizing the remarkable efficiency of AI in learning. The process unfolded in three stages. Initially, Claude curated a customized syllabus based on Fernández-Villaverde's prior knowledge. Following this, he engaged with Goffman's influential texts, including "The Presentation of Self in Everyday Life." While Claude structured the readings, he emphasized that he conducted the reading himself. Lastly, he interacted with AI as a learning partner, seeking clarifications and drawing connections to his existing understanding of economics. "It feels like having an infinitely patient colleague by your side," he remarked. However, despite its strengths, Claude has limitations. According to Fernández-Villaverde, AI does not challenge students in the way that skilled educators do, often responding to questions posed rather than prompting critical inquiries. It also lacks the collaborative dynamics of a classroom environment. Yet the professor acknowledges that human educators can struggle with accuracy as well. The implications of his experiment raise critical questions about the future of higher education. Many academics, including University of Texas history professor Steven Mintz, share concerns that AI is exposing systemic weaknesses in traditional educational models, highlighting issues such as standardized assignments and outdated teaching practices. Fernández-Villaverde argues that AI poses economic challenges for universities, compelling them to justify tuition costs. Institutions primarily focused on delivering lectures may find their business models under severe scrutiny, especially when students can access comparable knowledge for a fraction of the cost. Despite these challenges, he does not foresee AI eliminating traditional education. Instead, he predicts it will prompt universities to innovate and offer unique value that AI cannot replicate, such as research mentorship, access to laboratories, and robust peer networks. "The institutions that will succeed are those that provide what AI cannot," he concluded, emphasizing the need for higher education to evolve.