Does AI Mean Never Having To Say You’re Sorry?
First published: October 20, 2025. Posted here for archival purposes.
“Everything in Between” is about the systems, institutions, and practices that people build, “things” of a sort that sit in between us, between groups of us, between “us” and “them,” and between us and other systems and institutions that seem terribly far away: “the market,” “the state,” the universe, and so on.
New for me
What if Generative AI is an enormous MacGuffin relative to the continuing evolution of modern universities?
That question borrows a Hitchcockian phrase for an object with little meaning in itself that nevertheless drives a movie plot. The glowing briefcase in Pulp Fiction, for example.
If colleges and universities have plots, then those plots usually go something like this:
Teaching in a college or university environment involves a delicate and nuanced dance among relying on your expertise to guide, inform, and perhaps even inspire your students; listening to your students, who may have plenty of wisdom of their own to share; and creating conditions where students share with and learn from one another, and not only from their teacher(s).
Those are processes and systems with interwoven feedback mechanisms and imprecise techniques for evaluating successes, failures, and opportunities for revision. Their connections to intellectual, social, and/or emotional content, let alone to intellectual or other skills and competencies, is incalculable, meaning that it cannot be calculated, not that we do not know, in some respects, that those connections exist (at times) or don’t (at other times). Teaching and learning are, almost inherently, inefficient in that sense. To be a bit more precise, we cannot know well, consistently and accurately, when they are efficient and when they are not. Not knowing is part of the point of higher education systems (and perhaps all education systems), rather than part of their failing, even though everyone, teacher and student alike, also knows that time, money, and expertise are precious resources and should be produced and curated with great care. Universities are knowing and not knowing, simultaneously, top to bottom.
Now the plot thickens:
Into that inefficient mix we have thrust Generative AI systems. Or, as many faculty and administrators might say, Generative AI systems have been thrust upon us – perhaps by vendors of educational technology (EdTech), perhaps by BigTech, perhaps by consultants, perhaps by central IT managers, but perhaps most of all by students. Hollis Robbins is quite right when she observes that university faculty may be fairly characterized, collectively, as steering curricular and pedagogical horses and buggies in environments where our students are rapidly whizzing by in metaphorical automobiles of ever-increasingly sophistication. And she is also right to point to university leadership (well, management, in any event) for their collectively slow response to the challenges of updating the university’s 19thc and early 20th century epistemological and governance conditions for student, faculty, and institutional success in this technologically-transformed world.
To sum up and switch the metaphor, faculty are, in a way, losing the plot.
I have three observations to add, suggesting that Generative AI is not quite the character that it seems to be in this story:
One goes to cars and their effects. As the historian Peter Norton has written, the introduction of the automobile produced a large-scale revision to society’s collective understanding of what streets were (and are) for, and of who streets were (and are) for. The new framing was (and is) this: streets are for cars, rather than for people. Walking in a street without permission is criminal, although that was not generally true before cars came to be. Translated, my point is: the university is for some people and purposes and not for others. That has long been true. When we add Generative AI, whether it comes from university IT management or from our students, the “who’s in and why?, and who’s out and why?” is the meaningful question that should drive curricular and pedagogical updating.
Two goes to institutional framings. An enormous amount of the broad public (and also specific local) conversation about Generative AI in higher education has come to rest, for now, on the shoulders of individual teachers and individual students. Teachers are expected to learn what they might about AI, to develop and implement policies for their students about AI, and to enforce those policies. Students, too, are expected to listen to and adhere to individual teacher policies, recognizing all the while that policies may vary widely from teacher to teacher.
In a way, this is ordinary academic freedom at work; few teachers at any level would agree that it is a good idea for them or for students to have a “one ring to rule them all” Lord of the Rings-style mandate from a central administrator, controlling how AI is used for teaching and learning across all classrooms, fields, disciplines, and schools. (There is a MacGuffin reference if there ever was one). At the same time, the absence of mechanisms for coordination, conversation, and reflection across teachers, fields, disciplines, and schools, including how teachers learn from student investments in AI, is deeply problematic. (Hollis Robbins has a different take on the common “silo” metaphor for disciplines in higher education.)
Again, the meaningful question is not simply (for example) “how can my law school borrow AI-relevant expertise from the department of computer science?,” or the reverse, although those are important questions. Instead, the meaningful question is: “how should communication and coordination mechanisms be deepened and made more effective across the full range of the university, including disciplines and fields, for all purposes?” That leaves me agnostic as to the strengths of existing disciplines; I focus attention on layers of substantive connectivity that should enable productive interweaving.
Three goes to the intersection of culture and incentives. Years ago, when I first moved to Pittsburgh, early conversations with people in the region’s then-very-small-scale business and tech sectors yielded a collective conclusion: in the late 1990s, Pittsburgh was still experiencing the harms of the collective, 15-year-old psychic wound inflicted by the collapse of the steel industry in the region in 1982 and 1983. Entrepreneurial activity and small business formation in Pittsburgh was at pitifully low levels; virtually all of the public sphere conversations about renewal and economic development had a deeply nostalgic character. How could Pittsburgh re-build large-scale integrated industrial organizations? Big business largely left Pittsburgh; the only route to renewal lay through rebuilding big business.
My new friends and I diagnosed the problem as cultural, not purely economic. Investment capital was available, at least in principle; public policy was not, itself, a barrier to new economic activity. Material incentives, in other words, had not helped the region advance, although they may have slowed the region’s decline. Looking more broadly and ambitiously, we asked ourselves, how can a city change its culture? Its personality? Its collective ambition?
The answer that I’ve come to, after 25 years of pondering, listening, observing, and occasionally writing, is that you can’t. If the culture needs to be changed (meaning: if the design and operation of the institutions inside the institution of the city need to be updated), then current or incumbent leadership and management can’t do it. (Maybe trying to change that culture is an error in the first place; maybe the community does not want to change, or objectively speaking, should not change. I leave those possibilities for a different time.) Current or incumbent leadership and management are too wedded to the old systems, by virtue of how they were educated, trained, promoted, and rewarded, to invest in the new.
Instead, the city has to experience the occasional chaos metaphorically borrowed from the car chase across the streets of San Francisco that appears in the image that begins this post: While everyone in the picture chases the MacGuffin (a red suitcase; growth; progress; efficiency; what to do about Generative AI?), incumbents leave and pass away; new people come in with different ways of thinking and doing things, occasionally breaking the rules (and other things) but over time laying down the foundations for a new order, partly bottom-up, partly top-down.
(The concluding scene of “What’s Up, Doc?,” features Ryan O’Neal, love struck, turning to Barbra Streisand, the Hepburn-esque agent of chaos in the movie, and quoting the line that he made famous in “Love Story”: “Love means never having to say you’re sorry.” Streisand levels him and closes the romantic coupling with her reply: “That’s the dumbest thing I ever heard.”)
Streisand is wise; institutional design (life) is not a romantic comedy. To the extent that Pittsburgh today has a reputation for being at least somewhat forward-leaning in its economic development potential, owing to what we used to call “eds and meds” investments and what we characterize today in terms of robotics and AI research and development, the cultural energy and investment capital mobilized in those directions come almost entirely from a “new guard” of regional leadership, slowly and incompletely nosing the “old guard” out of the way.
In short, instead of (or perhaps in addition to) “what should higher ed do about AI?,” we might ask: where is higher education’s new guard?
What others have to say
Selections from new-ish commentary about the institutions of higher education and institutions of expertise, shared because these folks have me thinking, not necessarily because I agree.
Hollis Robins at Anecdotal Value, “Token-Relief”:
NVIDIA’s “under the hood are tokens” is doing something weirder even than Marx. Look under the hood, they invite you, and you’ll understand what AI can do, the social relations going forward: the weather forecasting, the medical breakthroughs, the helping people walk and talk again. All good! But still invisible are the social relations that produced the tokens: the training data, the labor that produced it, the decisions embedded in the model, the infrastructure running it. Those faces don’t appear on any token.
(Tokens are AI MacGuffins! Is AI a series of MacGuffins all the way down? As for invisible social relations, the scale and complexity of the re-regulatory problems and institutional design problems associated with the introduction of the automobile is well-documented in this paper by the late J. Willard Hurst, a legendary law professor at the University of Wisconsin. The paper was drafted initially in 1949 and excavated and finally published only in 2022, thanks to the efforts of current Wisconsin law professor B.J. Ard and University of Michigan historian William Novak.)
David Hummels and Jay Akridge at Finding Equilbrium, “Economic Growth, Social Value, and Capacity Building: The Value of University Research: Part II”:
Because knowledge diffuses broadly and unpredictably it can be challenging to generate a proper counterfactual where you could say: without university research, what innovations would not have happened? But there are interesting thought experiments that drive at precisely this idea.
In this post we discuss evidence on how these indirect benefits contribute to economic growth, and to social value more broadly defined. We’ll also make an argument for how engaging in research raises the value of the university itself, even when that research has no obvious use case.
(I like to imagine that Joel Mokyr would approve of their argument.)
Kwame Twumasi-Ankrah at Pattern of Play, “The Self-Regulating Pitch: League tables miss what the streets teach”:
The league table isn’t neutral; it comes with cultural messages about what’s important in football. In a formal environment, position, points and goal difference have been established as the key criteria. But this kind of measurement reflects particular cultural priorities around time, competition and success.
The international street football landscape exposes different cultural practices to measurement of performance. As research on talent development has emphasized, “the magic of street football lies not merely in the format, but in the absence of authority, the risk and the permission to fail and game on. You can’t plan, coach or track it. What you can do is make room to do and let it happen.”
This isn’t about throwing away formal leagues, but that the success of various cultural settings is measured to achieve a different criteria. In street football, success isn’t about final position; it’s about adaptability. And when players around the world cycles through the “winner stays” scheme, they aren’t vying for points but for growth.
(I was delighted to discover this newsletter, which features writing about soccer (football) that looks at the institutional and systems attributes of sport.)
My bookshelf
Like a lot of academics, I read a lot. Like a lot of law professors, I read a lot about law and about governance. But I also read a lot of things just for fun and a lot of other things because you never know where interesting ideas might come from.
I am still in the middle of this book, which I teased last week:
Mark Blyth, Great Transformations: Economic Ideas and Institutional Change in the Twentieth Century (Cambridge University Press, 2002). The publisher’s description:
This book picks up where Karl Polanyi’s study of economic and political change left off. Building upon Polanyi’s conception of the double movement, Blyth analyzes the two periods of deep seated institutional change that characterized the twentieth century: the 1930s and the 1970s. Blyth views both sets of changes as part of the same dynamic. In the 1930s labor reacted against the exigencies of the market and demanded state action to mitigate the market’s effects by ‘embedding liberalism.’ In the 1970s, those who benefited least from such ‘embedding’ institutions, namely business, reacted against these constraints and sought to overturn that institutional order. Blyth demonstrates the critical role economic ideas played in making institutional change possible. Great Transformations rethinks the relationship between uncertainty, ideas, and interests, achieving profound new insights on how, and under what conditions, institutional change takes place.
Next up after that: Sarah A. Seo, “Policing the Open Road: How Cars Transformed American Freedom” (Harvard University Press, 2019).
Thanks for sticking with me.