This document was written prior to the TFUAP committee’s updated report and FAQ, which respond to a number of the topics discussed below. We appreciate the committee’s engagement with these concerns, but believe further discussion remains warranted. 

Thesis: This letter highlights the intellectual risks and pedagogical challenges associated with the proposal by the Task Force on the Undergraduate Academic Program (TFUAP) to compress science General Institute Requirements and suggests an alternative framework to infuse concepts of probability, statistics, and machine learning (PSM) through Institute-wide collaboration inspired by the Common Ground framework. We employ curriculum of the Department of Materials Science and Engineering as a test case to demonstrate the impact of the TFUAP proposal and the intellectual benefits of the collaborative approach to ensuring PSM fluency in MIT graduates.

This letter is presented to the Task Force on the Undergraduate Academic Program (TFUAP) in consideration of the proposed changes to the undergraduate education at MIT. We appreciate the extensive work of the TFUAP, and the thorough nature with which all aspects of the curriculum have been examined. We share the TFUAP’s goal of ensuring that MIT graduates are prepared to lead in a rapidly evolving scientific and technological landscape. We support the goal of incorporating computation, statistical thinking, and data literacy more explicitly in the undergraduate curriculum. A periodic review of MIT’s educational structure is vital to maintaining our leadership in science and engineering education, and the TFUAP report clearly shows dedication to this important goal.

Nevertheless, we believe the proposed restructuring of the General Institute Requirements (GIRs) is ill advised.[1] For decades, the GIRs have served as a shared intellectual foundation for all MIT students. These subjects establish a common baseline of preparation that allows departments to build advanced curricula with confidence in students’ prior knowledge. They ensure that all students are exposed to critical concepts from fields outside of those explicitly required for their chosen subject focus. The proposed shorter “exposure” subjects and integrated GIR options will fragment that shared foundation. While curricular flexibility has clear advantages, it may also result in significantly more heterogeneous preparation among students entering advanced coursework.

This heterogeneity will shift the burden of maintaining foundational training from Institute-wide GIRs to individual departments. Many programs depend on specific GIR subjects as prerequisites for advanced instruction. If students arrive with varying levels of exposure to core topics, departments would be compelled to add new prerequisites, redesign existing courses, or create “bridge” or catch-up modules to ensure students reach the necessary level of preparation. The proposal itself anticipates the possibility that departments may need to develop such catch-up options to cover gaps between different GIR pathways. As a result, the desired result of “returning units to departments” will be negated by the need to include these new classes as prerequisites, and the net effect would result not in an increased but a decreased flexibility in major design for many disciplines.

Using the undergraduate curriculum for the Department of Materials Science and Engineering (DMSE, Course 3) as a test case, there are concrete examples of these challenges:

• Loss of foundational preparation: The introduction of 6-and-6-unit pairings of some GIR requirements and the consolidation of the Physics GIRs into a single required subject could result in students entering the DMSE core without prior exposure to critical concepts that underlie many areas of materials science and engineering (semiconductor devices, polymer synthesis, photonics, etc.).
• Increased instructional complexity: Addressing emerging gaps would require additional subjects and/or the creation of supplemental 6-unit modules, increasing instructional complexity via the introduction of two sequential modules intended to reconstruct the depth of a 12-unit course.
• Diminished intellectual depth: Two separate modules that introduce material at increasing levels of depth are not pedagogically equivalent to a single coherent 12-unit course designed as an integrated progression of concepts. Introductory subjects are typically structured so that ideas are introduced, reinforced, and revisited in a logical sequence. Splitting this progression into separate exposure and catch-up modules disrupts that structure and will likely require substantially more than the nominal difference in units to restore the same conceptual foundation.

Collectively the proposed changes will sacrifice flexibility at the sophomore, junior, senior levels in Course 3/3A degree programs for flexibility in the first year. We argue that flexibility in later academic years is of greater value, because this is when students are best prepared and motivated to pursue opportunities (internships, ambitious research etc.) that are a springboard for their careers.

These challenges are not unique to DMSE/ Course 3/3A. Many departments rely on the GIRs to ensure rigorous preparation. Introducing multiple preparation pathways will likely force a broader expansion of degree requirements, shifting foundational instruction from the GIR system into the majors. This would undermine one of the central goals of the proposal (reducing complexity in the undergraduate curriculum) while curtailing students’ ability to pursue ambitious and self-motivated goals in their 3^rd and 4^th years, as illustrated above.

This letter highlights the downside risk of creating flexibility within the GIR to enhance MIT students’ foundational exposure to computation, statistical thinking, and data literacy. However, this should not be interpreted as resistance to curriculum modernization.

Departments across the Institute are already integrating computational and data-driven approaches, in recognition that such modifications are essential to future-proof the education that we proudly deliver. In DMSE, subjects 3.029 and 3.C01[J] (taught jointly with the Schwarzman College of Computing and Courses 7, 10, and 20) incorporate machine learning and modern data analysis methods, and our faculty expertise in these areas continues to grow with strategic hiring in computational materials science. The question is not whether new competencies should be incorporated into the curriculum, but how best to do so while preserving the coherence and rigor that characterize an MIT undergraduate education.

In the spirit of collaborative brainstorming, we suggest an alternative mechanism to integrate foundational exposure to probability, statistics, and machine learning (PSM) for every MIT undergraduate without eliminating a GIR. Just as we have not eliminated or curtailed established GIRs to teach essential communication skills, we propose that PSM should be integrated by adapting (or reimagining, as needed) existing courses within each major. A PSM Intensive in the Major (PSM-IM) credit requirement would provide the necessary structure to ensure that all MIT students graduate with foundational understanding of PSM, and an exciting incentive for creativity and collaboration in teaching PSM across the Institute.

DMSE/Course 3/3A is well on its way to implementing such an approach. Our curriculum includes multiple laboratory components in which students produce data (computational as well as experimental). Recrafting these labs to integrate PSM topics and to add suitable deliverables will provide both foundational and working knowledge. Creating lab-based PSM-IM subjects is an opportunity to teach machine intelligence (e.g., relevant to AI-driven experimentation), which today is a frontier topic but soon will be an essential competence. We are hiring new faculty in this field and plan to have a leading AI researcher co-teach an experimental lab class. Further, our degree chart includes requirement 3.029 (Mathematics and Computational Thinking for Materials Scientists and Engineers). This course will be collaboratively reimagined via collaboration with the Common Ground framework to emphasize PSM.

We believe similar implementations could be designed within other departments. It is exciting to imagine an Institute-wide rollout of PSM education in the context of lab experiences (now including computational labs) as a form of mens et manus for the AI, as well as for our students. That said, the suggested mechanism does not hinge on lab subjects alone. The broader opportunity is to create an imperative for all undergraduate programs to modernize without necessarily creating entirely new subjects and additional subject requirements.

Drawing again from our experiences with the CI-M requirement, a PSM-IM requirement would necessarily create new education-focused collaborations as experts in PSM topics are engaged to co-teach PSM-IM subjects. This structure mirrors the successful Common Ground program and would yield broader intellectual impacts across the Institute, for instance accelerating the adaptation of AI for research programs. This set of benefits that would likely not accrue from offering PSM foundational education as a standalone GIR.

MIT’s General Institute Requirements embody a cohesive intellectual experience that informs and contextualizes each student’s specialized education in their chosen field. As we seek to adapt the curriculum to match current and future STEM and societal challenges, it is essential that we preserve this common foundation while thoughtfully integrating new areas of knowledge. We welcome continued dialogue with the TFUAP and the broader faculty community to ensure that the next evolution of MIT’s undergraduate program strengthens both disciplinary depth and the shared academic culture that defines an MIT education.

[1] The proposed changes would reduce the number of GIR subjects from 17 to 15 and replace the current science core with a new Science, Math, and Computing framework that includes flexible and integrated subject options across chemistry, biology, computation, and probability/statistics/machine learning. It also eliminates the REST and Institute Laboratory requirements, returning those units to departmental programs.

Back to top