Duke AI Report
And more basic questions about doing undergrad at an R1
Note: This essay does not represent the views of Duke University, but are instead my personal views, offered under academic freedom.
*******************************
The AI at Duke Steering Committee released a report to the Office of the Provost right before Spring Break, and I have just begun reading it. The Provost took some heat when he announced last year that all Duke students and most faculty would be provided a subscription ChatGPT account last year. Some faculty felt this was a pedagogical intrusion that needed more discussion. The Provost stated that he viewed it as being more about equity of access and I think his was a reasonable decision, that also included the announcement of the process that culminated in this report.
There were about 60 faculty who served on four committees, and the work appears to have been generative as one would expect given the topic and there are six “interconnected focus areas” for how Duke University is engaging AI that are listed below:
Research: Strengthening institutional capacity in foundational, translational and human-centered AI, while advancing AI’s trustworthy and responsible use.
Education and Workforce Development: Expanding AI education, curriculum development and experiential learning pathways to prepare students, faculty and staff for success in the wake of technological transformation.
Communications and Thought Leadership: Developing a strong institutional identity around AI and elevating the university as a trusted leader in conversations about the future of AI.
Governance and Culture: Establishing durable and responsible AI governance frameworks to guide the ethical use of AI across research, teaching and university operations.
Infrastructure: Building the technical and human systems needed to enable AI research and innovation at scale.
Sustainability and Societal Impact: Aligning AI advancement with Duke’s commitments to environmental sustainability and social responsibility.
I would like to thank my faculty colleagues for serving in this way, and thus helping Duke to view the challenges through the lens of the primary work of research and teaching that is the reason a university exists (and health care delivery, and sports, and getting outsize returns from the endowment to pay for stuff, etc.). But at its bedrock, Duke is an R1 research university, and I like how the report integrates research and teaching questions into an interconnected whole. It is messy because everything about higher education is messy right now, and especially at a complex R1 with a health system and big-time sports.
You can miss the forest for the trees given all the noise.
Yesterday in a public policy faculty meeting, we were discussing the implementation of a simplified set of faculty titles and a new effort to encourage teaching excellence outlined as “university guidance with school flexibility.” A discussion centered on minutiae about titles ensued, but a colleague stated a question so clearly that it made my head snap up:
Aren’t we still a Research University? How much reduction in research time should we invest in teaching?
I have been thinking about a related question while working on a book. It is orthogonal to this question, or maybe it is not.
Why would a student attend an R1 Research University like Duke to get an undergraduate degree? Can we make undergraduate education better because faculty are also engaged in research?
The historical answer to these questions went something like “faculty engaged in the cutting edge of their research field will be the best teachers because….” I think the answer kind of trails off. It could be, “because they bring the cutting edge scholarship into the classroom” but I have my doubts. Some great researchers are great teachers, but not all. And of course we do not really know how to assess teaching. The student survey is incomplete at best and is likely far worse, but it is unclear what to put in its stead. Hence, my colleagues question of tradeoffs.
Most undergraduate teaching tends to be general and basic, and most faculty research tends to be focused and specific. The most difficult class in the undergraduate public policy major at Duke to get filled is the Intro Class, PPS 155. That is because there is a light touch on so many topics that no one person could be “expert” in all of them. I am really trying to focus on how to make the fact that I am a researcher who teaches make a positive difference in how I teach that class next Spring. I have done that with two other classes this year, but it is new effort, in my 29th year on faculty at Duke. More later on that.
A memorable student from a few years back shared with me that her final college decision came down to Davidson, Duke and UNC. She was from North Carolina, so in one sense this is an obvious list of great universities from her home state. Leaving cost aside (she said that was not an issue for her family, but if it was UNC is far cheaper) these three choices are quite different from the experience of being an undergraduate. Davidson has fewer students than her high school did, and is a classic small liberal arts teaching college. Duke and UNC are more similar in terms of the faculty being employed by a research university which meant they had to address the trade off my colleague posed in yesterday’s faculty meeting (how much less research for better teaching?). Duke has a smaller student body (by a lot) as compared to UNC, so there are likely more opportunities to be engaged in research with faculty where I teach as compared to where I studied. And I have undergraduate members of my research teams all the time at Duke, but never worked on a faculty project when an undergrad at UNC in the 1980s. I was the kid in the back row with a baseball cap pulled down to shield my eyes.
However, the most important questions are not about class size, but instead whether, why and how a research university has, or can have, a comparative advantage in teaching undergraduates?
A related question that is knowable, but I certainly do not know the answer, is what proportion of the classes taught during an undergraduate career at a research university are taught by faculty whose professional life is a juggle between research and teaching, and how much is taught be those whose focus is on teaching? How does this compare at a private university v. a public flagship? I do not know the answers.
AI has not created these questions, they have been lurking about in the shadows, ignored at least by me. Hollis Robbins in particular has been key in my learning trajectory over the past year. However, now they are at least asked. And that is good. We should know the answer and be clear about it.
In my experience, excellence in teaching and excellence in research are strongly negatively correlated. That is: the top researchers are often terrible teachers, and the top teachers are rarely successful researchers.
How could it be otherwise? Excellence in **any** profession is a full-time job. You can't be a expert lawyer **and** a professional basketball player, any more than you can be a top chef and a leading surgeon. In part, this is a matter of time: your time is limited, so every minute that you spend on activity X is a minute that you didn't spend on activity Y. Good teaching demands a lot of time, and there's always room for improvement, so a truly professional teacher would be spending 100% of their available time on developing their teaching skills. On the other hand, research faculty are under huge pressures to churn out papers and harvest grant funding. Something has to give, and inevitably that means that people cut corners with their teaching.
Think about the statistics involved: truly excellent teachers are rare, just as truly excellent researchers are rare. If 1% of the general population is capable of becoming an excellent teacher, and 1% of the general population is capable of becoming an excellent teacher, then 0.01% of the general population is capable of excellence in both. Most researchers are mediocre teachers because most people are mediocre teachers.
The notion that research improves your teaching is transparently bullshit, and it's just a coping mechanism for researchers to justify phoning in their teaching responsibilities. It's easy to see that this is nonsense -- nobody is arguing that good teachers make better researchers. Why is it that doing research has such profound impact on the quality of teaching, but teaching is irrelevant to doing research? The asymmetry here indicates that people aren't serious when they claim that there is a connection between the two activities.
By the way, this is true outside of formal academia. I've had the opportunity to take music lessons with some very big-name professionals, and they usually don't add much value. On the other hand, the really great music teachers that I've had have often had undistinguished performing careers. Again, being a great performer and being a great teacher are totally different activities which demand very different skill sets, and very few people have both.
I like your questions about undergrad teaching at R1s, and have thought about this topic a lot. I'm the product of a small liberal arts college (I won't say which one, but 100% of alums Strongly Agree that it is The Best One), and have spent my entire life since graduation at various R1s. This includes my early faculty career, where I did do undergrad teaching, and my later years at other R1s mostly focused on grad ed admin with a tiny bit of grad student teaching.
I and several of my colleagues at my first faculty job were considered to be great teachers. This was certainly appreciated by the "paying customers" (aka students), though was otherwise not considered an important aspect of our value by anyone else evaluating us. But did being active scholars contribute to our great teaching? I think at least a little, though maybe not as much as one might expect and not necessarily because we brought cutting edge scholarship to our teaching. In fact I was rather intentionally focused not on the latest and greatest, but rather at presenting the arc of inquiry in my field (dev bio) starting with establishing the foundational concepts and then examining the evolution of the methodologies that produced historical knowledge and define contemporary practice. But, to be honest, I don't think I needed a concurrent extramurally funded research program to do that. I think that the exposure to the field and methodology during a strong PhD and a productive postdoc would have been sufficient. I suppose that having an active funded research program gave me both confidence and externally validated authority. Students did indicate that they took us more seriously than they took the adjuncts, even though they liked many of the adjuncts better.
As (generally) satisfied as I am with my own career, I will admit that I have yet to achieve the kind of satisfaction that I assumed I would get when I was a sophomore having epiphanies in Cell Bio and Organic Chem inspired by the almost uniformly outstanding teachers I had at my SLAC. I probably romanticized their lives in small northeastern towns with their 2-2 teaching loads, cozy offices, student-centered presence and availability, and their attempts to get some tiny grants to be able to provide us with the minimal research experience we needed to get in to good grad schools. My R1 life has had more structure and formality and an entirely different focus.
I had all of this in mind as a parent when my kid was looking at colleges. I tried to subtly steer her away from R1s. Funny thing is she ended up at a R2 that presented itself, somewhat accurately in terms of student life, as very undergrad focused, and then proudly announced their attainment of R1 status during her junior year which I guess somehow justifies the nearly $100K that they extract from us annually.
Also noting how similar the Duke AI report is to other university AI reports that I have read (and the one that I am in the process of contributing to as we speak..).