Statement on AI in Undergraduate Education

The expansion of AI’s capabilities is as significant a development for higher education as any we have seen in our lifetimes. The new tools present as yet unsolved challenges to our understanding of what constitutes original work, authorship, and authentic learning—and they have provoked a reexamination of how we perform our core functions of teaching, assessment, and research. 

Whatever the shape of our future with AI, our society will need as urgently as ever those with the habits of mind that a broad education across the arts and sciences uniquely fosters. We need independent thinkers who understand the importance of context, who develop the acuity of judgement to be able to ask better questions, who can separate strong arguments from weak ones and fair from unfair uses of data, and who can connect and learn across different points of view.  

The shifting context in which we pursue these goals requires adaptation in the classroom. Since AI can assume some of the effort that’s required from us to perform intellectual work, it is up to us to answer which kinds of effort constitute what we might call cognitive “resistance training,” necessary for a great education, and which can be refocused to deepen learning. 

Given the heterogeneity of disciplines in our school, our approaches will be multiple. Faculty with expertise in their materials are best positioned to carry out coursework redesign across our departments and programs, with support from those with expertise in AI and pedagogy. AI will also have different roles to play at different times in a student’s College pathway, with an emphasis on ensuring resistance training to develop fundamental skills in the early years, and more possibilities for AI to deepen learning as students work closely with faculty in the major. 

Today, the College is focusing on four key areas: 

• rearticulating our shared principles of learning 
• adapting and innovating our pedagogy 
• expanding tech-free learning opportunities 
• making AI itself an object of study 

Our work in these areas aims to help students and faculty decide how to use, or not use, AI to strengthen what is essential in an arts and sciences education. The initiatives are summarized below.  

I. Recommitment to shared principles and practices in the classroom 

Successful adaptations in teaching will flow from our goals, expectations, and norms for student learning. Mindful of the rapid changes on many fronts at this historical moment, including the rise of AI, the Committee on Undergraduate Education (CUE), the College’s chief faculty governance group, is writing our Classroom Charter. It rearticulates the principles behind our commitment to residential, in-person, dialogue-based learning. It centers on these core elements: focus, free expression, trust, dialogue, and integrity and transparency in evaluation. CUE’s Classroom Charter will be a basis for faculty to make needed adaptations to course policies and the best practices that support them, including around AI.  

II. Pedagogical adaptation and innovation 

The College is working through our Undergraduate Council, standing committees, and plenary sessions with SAS faculty to raise up and share out best practices. On the side of mitigation of harmful uses of AI, these include assessments that measure process over product; re-centering paper-based reading and problem sets; using software that makes a student’s writing process more transparent; and an increased focus on in-class work, including reading and writing assignments, quizzes, and written and oral exams. In AI adoption, these include AI tutors trained on course materials; attentiveness to how large language models operate, exploring model biases, identifying common errors, and strategizing best querying practices; and guided use of AI in constructing and critiquing arguments.  

The College’s first-year curriculum incorporates our best thinking regarding effective student-facing AI policies, to be updated each semester. The new Foundations courses, Kite and Key, required of all first-year students beginning in 2027–28, are native born to an AI-infused world. Beginning this year, they have included tech-free study in co-required labs, instruction in the fundamentals of how AI works, an assessment structure built to mitigate unhelpful uses of AI. The Critical Writing Program has piloted low-tech and no-tech seminars, as well as those that use AI. The First-Year Seminar Committee is coordinating and sharing best practices regarding AI in their guidance to faculty. 

We have partnered with the Data Driven Discovery Initiative (DDDI) to develop a website with direct guidance for students on best practices around AI for learning. We are working with DDDI and CETLI on faculty-to-faculty instruction and workshops; department-specific channels for exploring course and assessment redesign; and a mini-course, “How AI Works,” to be available in hybrid form each teaching semester, for SAS faculty. The College also supports the Penn AI Pedagogy Initiative, led by the Graduate School of Education, which pairs interested faculty with a learning sciences researcher and a student technical partner to co-design AI enhanced coursework, test it, and contribute results to a shared repository open to educators worldwide.  

III. Tech-free learning 

To assess student and faculty desire for independent, AI-free workspace, the College launched a pilot in Fall 2026. We are offering monitored, device-free study areas during evenings and weekends, both to individual students participating on a voluntary drop-in basis, as well as to faculty who may assign some hours per week connected to their courses. This is meant to be a space to support those desiring distraction-free study and to ensure completion of regular coursework without AI, and not an alternative venue for exam proctoring. We will continue to work with campus partners to expand opportunities as appropriate.  

As indicated above, the first-year required courses Kite and Key also incorporate strong tech-free and deliberate-use policies. Faculty committees governing the first-year Critical Writing Seminar and First-Year Seminar are also carefully examining policies and means to facilitate tech-free coursework and class time. 

IV. AI as an object of study 

Better understanding of AI, including how it arrives at outputs, how it was developed, and what it can and cannot do, is a prerequisite for preparing our students to engage with AI not just as a provider of answers but as a subject of further critical engagement. Such understanding will allow students to adopt it more competently and judiciously, and to be more confident in their own capabilities relative to it. The College has committed to such teaching, including support for a new half-credit course, COLL 0201, “Introduction to AI: Applications and Impact;” inclusion of modules about AI in College Foundations; and support for an AI focus within the Data Science and Analytics minor. We are working with departments and programs on additional courses, including First-Year Seminars, that promote critical and technical understanding of AI.