In the Age of AI, Are Universities Doomed?

I n 2019, my friend Brian Cantwell Smith published a book about the power and limitations of artificial intelligence. We debated whether his title should be The Threat of AI, The Promise of AI, or the anodyne The Future of AI. He chose The Promise of Artificial Intelligence: Reckoning and Judgment.

When I was celebrating at a birthday party with Brian, shortly after the 2016 US election, I met Geoffrey Hinton, one of AI’s pioneers. He explained that Brexit and Donald Trump owed their victories to Cambridge Analytica and that, as a result, many people would suffer and even die. You don’t need me to survey the debate, primarily apocalyptic but occasionally hopeful, about the future of a society in which data is mined and trolls attack anyone they disagree with. The ability to disseminate and tailor news to the reader, to go well beyond digital manipulation to produce false images and false videos—this all seems to challenge our basic sense of knowledge and the task of learning how to discern truth.

New alarms were set off in 2023 with the rise of chatbots and their capacity to generate academic writing on any topic within seconds. I’m no expert on information technology, much less AI, but AI prompts me, as a philosopher, to ask: How can universities address this crisis about what will count as knowledge in the future? Or, more precisely: What can our students learn to guide them in this muddled storm so they can help society?

The university was seen for centuries as the repository for and guardian of information—higher education meant higher access to information. Information is now massively accessible. If the university exists only to warehouse knowledge and to transmit tradition, then it is doomed. But the university’s inability to claim a near monopoly on knowledge doesn’t make it obsolete. The focus must now shift to critique. The university needs to replace the question of access to information with an emphasis on judgment and interpretation. Such a shift, however, does not often enough lead to the idea of inquiry. Which prompts the question: What will research look like in the new information age?

Although the IT revolution seemed, at first, positive for the dissemination of knowledge, it has marked losses: loss of the goal of one unified body of settled knowledge and loss of a grand narrative justifying the university as a place where all knowledge can connect into a greater whole. In the face of the collapse of grand narratives of progress, many in the university are striving to restore one or another of those narratives to secure a future of greater freedom, greater productivity, or greater technological power. In contrast, most humanities scholars are interrogating and inquiring, asking questions with no expectation of a settled answer.

We live in a time of massive problems, from climate change to systemic racism, from the challenges of immigration to injustice, from pandemics to the flood of information. We need new solutions, new insights, new policies, new technology, but we have little time. The future is leaning hard on today. The university is a venerable institution, but if it can’t help us now, then it has no future.

• I Used to Teach Students. Now I Catch ChatGPT Cheats
• How ChatGPT Is Putting College Ghostwriters Out of Work
• Will ChatGPT Kill the Student Essay? Universities Aren’t Ready for the Answer

O ur current change in information technology requires a transformation in how scholars read and write. It’s clear to me that while students are out in front, traditions have much to offer.

When AI shifted from being something imagined as futural to something confronting universities—and all of society—concern about the place of essay writing in the humanities gained a sharper edge. To take a step back from the pressing moment, I want to explore older humanities practices that are valuable to the new environment. The humanities have the distinctive asset of a longer tradition: an awareness of the capacity to change and to reconfigure past and present practices.

When I began directing and planning the Jackman Humanities Institute at the University of Toronto as its inaugural director in 2008, several advisers encouraged me to look into the digital humanities. I could write at length about our various attempts, mostly failed, to find a way to support digital humanities at the University of Toronto—which is ironic, given the University of Toronto was a pioneer in the field over thirty years ago. Some disciplines were resistant to changing their methods along with new technology; others were trailblazing.

In any case, we commissioned a census of research professors and their projects and discovered seventy professors were funded to do digital humanities. Scholars were engaged in diverse projects: assembling databases of archaeological findings, artworks, or historical languages; creating dictionaries; mapping communications or scholarship networks; developing visualizations and augmented reality; and hosting digital colaboratories to engage students and other universities in explorations of medieval text culture or contemporary issues in Indigenous studies. Many of our most creative scholars were engaging with the new technology to pursue new research.

We created a research community called the Digital Humanities Network. It grew rapidly to more than 220 members, including 150 research professors engaged in over 100 projects. The Jackman Humanities Institute didn’t fund or coordinate this sprawling set of activities and research; we hosted a network through workshops, conferences, and pilot fellowships. We saw that graduate students and sometimes undergraduates were taking the lead on these activities, and links with computer science and other sciences led the institute beyond the humanities.

One project (or, rather, a pre-proposal or call for a proof of concept by a group of readers) in which I was involved illustrated new thinking about how long-standing practices in the humanities might preview the future of reading and writing.

Scriptural Reasoning is a community of scholars that has been meeting for over twenty-five years, primarily at Cambridge University and the University of Virginia but also with key participation at the University of Toronto. Readers from the Christian, Islamic, and Jewish religious traditions meet in small groups and read parallel texts from the three traditions, paying careful attention to the complex meanings in each tradition. As scholars, we aren’t seeking common ground or trying to convert one another; we are exploring other ways of reading and deepening our understanding of our own traditions. This kind of study relates to the idea of discord: the goal is not to find a deeper unity but to seek new paths to bridge gaps between traditions.

Some time ago, in conversations with some software developers, a few of us imagined an online form for this activity. We were joining in a quest for what I called “the Holy Grail of digital humanities”: a flexible online commentary platform. The obstacles were numerous: multiple languages, alphabets, canonical texts, and, of course, willing participants. (In the intervening years, new platforms for commentary have emerged—for instance, Perusall. But the goal of creating an adequate platform still animates various digital humanities projects.)

In the process of imagining an online model—a model we did not succeed in building—we recognized three distinct practices: collecting, writing commentary, and editing. The first would be the practices of selecting what belongs in the library; the second would be practices of responding to specific texts or published matters; and the third would be sorting, reducing, and consolidating the creative commentary work in order to publish new texts.

First, we would need to create a common library of the editions we wanted to cite and comment on, and we needed to choose which translations we would use. The library would need to include the traditional commentaries, the lexica, concordances, and some scholarly materials (old and new). The initial question was not how to digitize materials but what should be collected.

Second, we wanted a safe, semi-private (or is it “small public”?) space for registering our comments and responding to one another’s. In real time, and over time, we wanted a wiki-like conversation, or something of the sort, where our group and maybe other parallel groups could read together and write our thoughts, large and small. Comments might be on a specific word and its history, a larger theme (for example, diverse written versions of laws or stories), or our own commenting interactions.

Third, at regular intervals, we wanted a smaller group of scholars to edit the commentaries to present their fruit to a wider public. That fruit might be a single dramatic new insight, a debate, a new reading, or a program of questions that emerged from the comments. This eventual, distilled publication would create space for a freer, more expansive, and experimental exchange among the group of commentators. The knowledge that later editors would refine the work would remove the pressure of finality, allowing discussions in this earlier stage to unfold more openly.

As I indicated, there are many scholars, labs, and groups working on developing commentary platforms. My point is that these three practices have been long-standing in the humanities. They go back to the humanists of the early modern era and beyond. They can, however, be revitalized in new digital environments, and humanities scholars are the ones most likely to make direct contributions to deepening the reading and writing of the future.

I n a truly fascinating way, university libraries are not only reacting to changes in reading but are leading the way. Librarians model and support old and new practices. Libraries are not simply collections but also sites for discussions and places that assist and guide collaborative research.

Thus, in terms of collecting or assembling the library, research in the future will need to sift through and privilege better sources. Google Books, for instance, is an ever-improving hodgepodge of virtual books. To reflect critically, to explore the limits of existing information, or to seek new knowledge, a person would need to learn how to assemble a relevant data set. In the humanities, for instance, there can be different critical editions (thank you, humanists) of a given text. Scholars can assess which are best and base their readings on the superior editions.

But as much as we need to assemble the best library, we also need to learn how to evaluate the vast array of information available online. Beyond searching using Google or Google Scholar or DuckDuckGo, we need to judge what should be included. An educated person can learn how to sort, prioritize, and exclude from the flood of information available to determine which of the old canonical texts and dictionaries are best and which sources for researching contemporary culture are dependable.

Commenting on texts is the ancestor of blogging. The ability to repost, offer my views, and criticize what others have said about another person’s post is now unrestricted. But what makes a good comment? Is the citation index (the number of times others cite something) the measure of a good comment? Perhaps obscurity or minimal citation is a mark of wisdom, of the rarer insight. While many great commentators in the past were individuals, they were rarely solitary. They shared their comments with one another, and their readings were sometimes preserved. Those early commentary pages are the equivalent of having multiple windows open in a browser. Reading is finding a way through these sources, and commenting is adding our own voices.

Finally, while tools like wiki editors, meta blogs, and emerging AI tools help consolidate the vast array of communications on the web, the practice of redaction remains uniquely valuable. Selecting the best materials from our group’s collective efforts—deciding what to share to advance research and knowledge—requires distinctive skills. While AI can use algorithms to sort and digest, and bots can proliferate provocative comments, the kinds of thinking that advance inquiry descend naturally from humanistic research.

Humanities scholars serve as editors of journals, book series, and edited works, and most review other people’s work for publication—our disciplines cultivate judgment on what is worth sharing. Can we teach this in the new context so that people can learn not only from ready experience but also from reflective and thoughtful engagement with knowledge? Can we revise the work of humanities scholars to educate students on better practices for thinking as they navigate the online world?

The task of the university in our digitally mediated world is more vast than collecting, commenting, and editing, and there are many other ways to conduct data analysis or learn how to navigate information online. But reading and writing are now unremitting. In earlier generations, small groups of people wrote letters, constantly. This is, in fact, a topic of study among humanities scholars. In our time, everyone is constantly reading and writing—through email, Facebook, TikTok, X, Instagram, and so on. We are caught up in a semiotic exchange. Media literacy is a learning outcome in many schools. Universities have a unique responsibility and capacity for teaching the kinds of thinking that will deepen and transform widely accessible information.

One of the new tasks for the humanities, specifically, is to revise and recast our traditions of scholarship for this new context. If you’ll forgive the pun, we should take a page from the humanists. We need to design new protocols to engage our students’ insatiable desire to read and write in these new environments so they will learn how to reflect and search for new knowledge. The challenge is speed. The current reading pace is fast—even though the time it takes a person to read a whole text might be quite long because they’re constantly jumping to other windows, following chains of connection away from the original text.

The kind of reading and thinking and writing that scholars pursue takes time. Perhaps, in part, what university education needs to do is slow reading and writing down. We now have rapid access to sources—which means we have more time for thinking and questioning. But having more time for reflection, rereading, revising, and rewriting is in tension with current views on efficiency and productivity. Going slowly might yield greater long-term benefits. Fostering habits of reflection, exploration, and discernment that lead to more valuable comments—this is a task for the university.

Adapted and excerpted, with permission, from What Could a University Be? Revolutionary Ideas for the Future by Robert Gibbs, published by the University of British Columbia Press, 2025. All rights reserved.