Technology

Three Things Thursday: A Warm Take on Magnifica Humanitas

Like many people, I’ve been reading Pope Leo XIV’s first Encyclical, Magnifica Humanitas. It is long, it is expansive, and in many ways, it is daunting. Time will tell whether it is a worthy successor to Leo XIII’s Rerum Novarum (1891), but it certainly continues (explicitly) in that tradition. 

There are already quite a few “hot takes” from various commentators on this document and I am sure that folks much more learned in matters of theology, “Social Doctrine“, and papal politics will have more thoughtful takes on this over the next year or so. I offer here a few warm takes on the letter that mainly focused on parts of the text that resonated most deeply with me. To be clear, I’m not Catholic and don’t have any particular faith commitment to this document, but it still struck me as a remarkable statement that speaks to many issues that haunt our contemporary society.

Thing the First

Over the last few months, I lost my father and this week we decided that our 13 year old “yellow dog” is suffering too much. Fortunately, my dad liked dogs and because he was willing to give Milo “butt scratches,” Milo liked my dad. I’m sure that Milo and my dad will find each other in the next life. (This sentiment is probably not theologically justifiable, by the way!) 

One of the points that resonated with me the most in Magnifica Humanitas is how human interaction is not something that benefits from efficiency, optimization, or improvement. When my brother and I visited my dad before he passed, he was brought to tears that we came to see him and wanted to spend time with him. If you know anything about dogs (and particularly our “yellow dog”) most of what they want is companionship (and treats) whether this took the form of just being in the room, going for walks, or a vigorous game of “ram ball.”

The genuine experience of time together is not something that AI can simulate and it is not something that can be performed more efficiently. It is irreducibly human and forms the fabric of our society and extends at least as far as our companion species (although this is not something that the Pope discusses). For Leo XIV, technological developments cannot be just if they lead people to feel alienated, lonely, and isolated through either false promises, shallow simulations, or material requirements that promote slow violence that rends the social fabric of communities.

Thing the Second

Leo XIV evokes the idea of the dignity of work drawing heavily on Leo XIII’s thought (especially in Rerum Novarum, but elsewhere as well). I have to admit that I had not considered how this concept fit into my formulations of “slow” in archaeology (or in academia more broadly). On the one hand, I had recognized that efficiency often drove scholars to race to keep up with ever accelerating expectations and standards and that these often measured accomplishments on the basis of quantifiable products (publications, citations, FTEs, or whatever) rather than growth experienced through process. On the other hand, I had not thought about how this growing sense of manufactured urgency and the shift from process to product dehumanized work itself. 

Magnifica Humanitas got me thinking more about how the tools we use carry with them certain social expectations that erode the dignity of work by accelerating the time when social and moral reflection take place. Many tools — particularly AI — see process as something to be optimized rather than enjoyed, savored, or celebrated. Pope Leo XIV’s encyclical calls out this reasoning and makes clear the process is what imparts work with dignity and allows for growth by making work human. When humans do work at the pace of the machine or do work to support the machine for the sake of the machine, this deprives the work of dignity.

The most scathing insight in this document is that those responsible for creating a society where optimizing process deprives work of dignity are morally responsible for the results of their actions. It is not possible to hide behind “the market” or “capital” as an excuse for advancing anti-human ideas. As humans, we have a unique kind of agency and as a result, unique responsibilities.

Thing the Third

Part of the genius of Leo XIV’s sprawling encyclical is that its structure, content, and scope make it resistant to contemporary reading practices. Much like Augustine’s Confessions, this is not a text susceptible to executive summary or other forms of streamlined digestion. Running it through your average LLM powered AI bot overlooks the nuanced interplay between the various arguments that relies, in no small part, on understanding the development of the church’s social doctrine over the past 135 years. Indeed, the interplay between Rerum Novarum and Magnifica Humanitas alone as well as scripture, Augustine’s work, and the writing of Leo’s immediate predecessors ensures that this text is not reducible to a series of bullet points, but rather a gateway to further engagement. The text is not linear. It is discursive and held together by centuries-old conversations, debates, and texts. It’s the kind of text that will draw a reader back to it multiple times and resist our society’s need for efficient, tidy, and conclusive engagement. 

In this way, Leo’s text resists the very forces that he (and his predecessors) see at play in the world. In a world that celebrates efficiency often at the expense of humanity, Leo has provided us with a very inefficient, but also very human text. 

Teaching Thursday: I Learned it By Watching You

This year I’ve been enjoying the handwringing about the use of generative AI by students. It’s been exciting to see how people hardened into camps and how fierce the “debates” have become. I’ve even come to enjoy the sometimes cloying moralizing that characterizes the “Never AI” camp and the techno-utopian imaginings of the pro-AI camp. 

Of course, part of the reason why generative AI is marketed to students looking to avoid having to write papers, literature reviews, or other kinds of assignments is because these assignments are both highly formulaic and very common, it is easy for generative AI to mimic the structure, tone, and even content of these essays. This reflects the formulaic character both of academic writing and also of academic thinking. There are exceptions of course; scholars who can turn a literature review into a nuanced intellectual history. But we should be honest that most literature reviews, for example, are not the most valuable parts of the articles that we read. More than that, as scholars we often lean on things like critical book reviews and historiographic essays to help us unpack the relationships between works of scholarship and streamline our understanding of disciplinary practice. We can argue that book reviews and historiographic essays are still human generated, but their tendency toward formulaic expression and standardized organization make them a very constrained and, as a result, banal form of writing (in most cases). The line between this kind of writing and that gloop produced by generative AI tends to be fairly thin. As a result, when our students read our work and the work generated by large-language models, they often fail to discern the distinction between human-made and machine-made interventions. This is as much our fault as writers and thinkers and our students’ fault as readers and learners.

As pressure to publish or perish becomes ever more pointed, the tendency toward formulaic work becomes more pronounced and even necessary to keep the scholarship machine (and the scholarly publishing machine) humming. In short, we’ve created the perfect storm for generative AI, and the increase in articles written by robots demonstrates that scholars are not immune from the temptation to take shortcuts. This, in turn, feeds the proliferation of journals, ranking system, and impact factors. 

When we think about how our students engage with AI, it perhaps would behoove us to start with our own behaviors as scholars. Students pick up on our priorities. If we treat writing as a transactional activity designed to satisfy the requirements of funding organization, to fortify our impact factor, or to maintain our contraction obligations, our writing (research and thinking) habits will show this not only in what we write, but how we write (and teach writing).

This semester, I’m facilitating a faculty reading group on Tricia Bertram Gallant and David A. Rettinger, The Opposite of Cheating: Teaching for Integrity in the Age of AI. (2025).  I’ve blogged on the book here. 

So far, attendance at the reading group has been highly uneven (to be polite). Moreover, most in attendance haven’t completed the rather short reading. To be clear, these reading groups are voluntary. Presumably faculty and staff signed up for these groups because they anticipated some benefit. They had the option of several different meeting times and the meetings were scheduled weeks in advance. Still, faculty struggled to turn up, struggled to complete the readings, and struggled to participate.

Just as pressures on faculty to publish or perish have contributed to student use of AI by reducing the complexity and creativity of academic writing, faculty and students share innumerable competing pressures when it comes to attending class, doing the reading, and participating in discussion. I’ve written about the challenges of attendance on this blog a number of times (here, here, here, and here) and how it isn’t a sign that students (or for that matter, my colleagues) don’t care, but rather a sign that our expectations are increasingly incompatible with current realities. At a minimum, our own struggles with attendance should make us more able to empathize with our students.

At best, it should offer a kind of insight into why things like generative AI offer such an appealing short cut. When faculty struggle to find time to do all that they want to do, this creates conditions where the shortcuts promised by AI can thrive. Instead of meeting at a set time to discuss a book, we discuss its contents with an AI bot. Instead of doing the entire reading, we ask AI to summarize the text.  These conditions extend to our students as well (and to our administrators, our friends, and to our lives outside the university).

Teaching Tuesday: Alienation, Affect, and AI

This semester I’m teaching a lot. A lot. It means that I’m busy, but mostly busy in a good way. As I’ve discovered in the past, I’ve found that the more I teach, the more I think about teaching and the more I’m likely to invest in the little habits that lead to reflection.

This weekend, on a long bike ride, I thought about three things that are shaping my teaching experiences this semester.

Thing the First

For the first time in over 20 years, I’m teaching a class where I didn’t pick the readings. This class is in our Teagle Foundation funded Cornerstone program. The class is a one-credit class numbered English 100 that is accompanies the second course in our composition sequence (English 130, I think). I have the students for a month where we talk about a series of texts that get students to think about the experience of being educated and how it changes our relationships and attitudes.

The class is less important right now than my experience teaching texts with which I was unfamiliar and also didn’t quite understand how they would (or were intended) to work together. At first, I felt very alienated from the texts (and the class itself) and struggled to figure out what to say to stimulate conversation about them.

Over time, however, I began to appreciate the humbling experience of teaching unfamiliar texts and exploring them alongside the students in the class. Freed from some kind of commitment to a particular interpretation (or even clear relationship between the course’s learning goals and the texts), the class has opened up in new and unexpected ways. Students seem more willing to take risks especially as it becomes clear that I have no clear grasp of how to make sense of many of these texts. If nothing else, it freed me from my (more than) slight (cough) tendency to use the Socratic methods when a discussion heads in an expected (and perhaps unwelcome) direction. 

Thing the Second

The other thing that has come from this class is that it has encouraged me to think about how students come to engage with the humanities. Since I was not particular familiar with the texts and not particularly comfortable talking about contemporary fiction and essays, I could offer relatively little in way of methods or even interpretation. What I could invite the students to do is experience these texts not intellectually, but emotionally. In many ways, this class has become an exercise in creating space for affective learning.

This loosely follows my own experiences in college where I was drawn to the humanities not because I found the methods or interpretative frameworks or theories of knowledge compelling, but because I like the classes and texts. That was it. I just liked the texts and I enjoyed talking about them.

What this oddball 1-credit class is doing is giving students permission to enjoy learning.  On a campus increasingly driven by assessment, learning pathways, checklists, and pressure, we’ve created a space (perhaps accidentally?) where students can just hang out and read texts and talk about them. And they have permission to enjoy it.

Thing the Third

I’ve been pretty flippant about AI use in my classes not because I don’t care, but because I have the arrogant assumption that if I tell students not to use it and explain why, they’ll understand and agree.

So far my experiences have been mixed. On the one hand, students seem to not use AI to improve their capitalization, grammar, or the organization of their papers. On the other hand, there is a clear increase in fact dumps in even low-stake writing assignments. This is a bit of a surprise to me especially since Microsoft Co-Pilot is baked into the cloud-based Word that many students use.

I suppose that I’ll have to intervene again. 

Traveling with a Notebook

One of the Things That I Want To Do (but probably won’t) is to get into a better notebook discipline. I’m taking two notebooks on my trip to Korea: an A5 and B5 size Leuchtturm notebooks. My plan is to use the larger notebook to document my trip and carry the smaller one for quick notes (e.g. in a museum or when it would be impolite to pull out a larger notebook). 

Some of my preparation has been mental. I’ve been reading Orhan Pamuk’s Memories of Distant Mountains: Illustrated Notebooks, 2009-2022 (2024) as a way to inform my notebooking practice. Of course, I’m not ever going to be able to replicate Pamuk’s use of illustration, nor his poetic and reflective language, but perhaps my notebooks could encourage me to look around a bit more and try to be write about the present rather than in support of some vague future writing goal or plan. 

Between multiple days of travel when any writing should be difficult, I’m only anticipating three or four days (starting today!) where I have the bandwidth and time to write something. This should be manageable, even if it is just quick notes when I take a moment to jot down what I’m seeing and experiencing. 

This means, of course, attempting to press pause on a plan for NDQ, not becoming distracted by the looming semester, worrying a bit less about discussion protocols in my survey classes and how my first semester teaching the Cornerstone Program will proceed, and relaxing about other challenges in my life.

Jacques Ellul, AI, and Teaching

I made a classic mistake this week: I decided to start a 300 page book. It was impossible that I would finish it before the semester started to gain momentum and the waning days of my summer research and writing time would brusquely push aside any time (or honestly motivation) to read a book.

So, I can’t imagine finishing Nolen Gertz’s Nihilism and Technology (2nd Editing, 2024), and this is not because I’m not enjoying it and learning from it. I did, however, find time to read Gertz’s recent-ish piece in Commonweal on Jacques Ellul and AI. (Artificial Intelligence, not Allen Iverson, although that would be awesome). The piece is short enough and good enough that it’s worth just reading. I’ve used Ellul’s ideas in some of my writing in the past. 

The one thing that I took away from Gertz’s (and Ellul’s) argument is that using AI for writing assumes that the inefficiency in writing is a bug rather than a feature. This follows Ellul’s arguments that technology (and more broadly “techne”) has created and perpetuates the privileging of efficiency (and scalability) above and beyond all other goals. Ultimately, efficiency becomes a goal of its own and inefficient processes tend to attract technological solutions. Writing, which is inefficient for many reasons, was a natural fit for technology that aimed at producing greater efficiency. The recent growth of Large Language Model driven AI is hardly surprising. After all, who has the energy, time, and bandwidth, to write the dozens of cursory email that academics write every day?

That said, the reality is that most of my writing isn’t about producing a finished product (efficiently or otherwise). A quick read of this blog makes clear that my capacity to proofread, edit for style, and even articulate myself clearly remains a work in progress. I’ve started to write a few times a week in a notebook to create space for even more provisional writing, stuff that wouldn’t even necessarily have a place in a blog post.

Writing, then, for me is about thinking. It’s about process. And it’s about discipline. 

These are processes that resist efficiency in profound ways. There is no shortcut to the practice of writing 1000 words a day. You just have to do it. There is process to putting together thoughts in an orderly way on a consistent basis other than doing it over and over. And there is no short cut to the benefits that come from writing consistently which range from writing more easily (or at least enjoying writing more) to thinking more clearly (you’ll just have to trust me here!).

For my students, almost all of the writing that we do (99.9% of it) is provisional. None of the ideas that we articulate in class and in papers are meant to be the final word on any topic. What writing is meant to do is help students sharpen their thinking process. The same way reading helps students become better readers. Practicing an instrument helps a musician become a better player.

To circle back to Ellul, then, our job as historians (or as scholars in the humanities more broadly) is to make the argument that what we do and our students do isn’t an inefficient process grounded in antediluvian habits or values, but rather an integral part of the development of historical (and broadly humanistic) thinking. In other words, it’s a vital part of learning to think. 

It might be conspiratorial to observe that thinking is an inefficient process in and of itself. Most animals react more efficiently when they don’t need engage in thought and just react whether through instinct or training. (To be clear, I recognize that a trained or conditioned response does require some thought, but it’s not what we’d recognize as conscious thought.) That said, the inefficiency of thinking is what allow us to understand difficult questions, to address challenging problems, and to exercise discernment. AI for all its glibness with language has not proven particularly adept at framing or even answering difficult problems. And when it has produced valuable new insights, this is largely driven by human inputs. In other words, humans have done the work to formulate the questions, which reflects the capacity of human thinking to search for meaning and order in the world. 

Fortunately, most of my classes privileges the ability of writing to help us not only frame questions but to attempt to answer them. Since this is not the domain of AI — yet, and perhaps ever — it remains fairly easy to explain to my students why it is not a viable substitute for the challenging and inefficient work of writing in my class. 

Two Things Writing Tuesday

I’m back from traveling and my summer research leave and while I’ll take a few days to recover from jet lag, I’m looking forward to getting back into my routine and starting the slow ramp up to the fall semester. There are two (and a half) things that are “closest to the sled” right now and I want to get sorted before the end of the month.

Thing the Half

I’m so close to being done the final report on my work with Richard Rothaus and Scott Moore at Isthmia. I previewed some of this last week, but I’m excited to share the entire report when it’s done. In other words, stay tuned!

Thing the First

In the spring, I wrote a draft of a chapter of a volume on campus crises. For whatever reason I misunderstood the directions and wrote it at twice the length that the editors wanted. This means that I need to cut this thing down in a pretty substantial way.

Before I do that, though, I want to make the full length paper available (and link to it in my more abbreviated and presumably published version). Here’s where it stands at present. Stay tuned for a shorter, more concise, and hopefully more on point paper.

Thing the Second

I’m very excited to work with David Pettegrew on paper titled “Mobilizing the Archaeological Report for the Future Interpretive Community: Linked Open Data, Analysis, and Publication.” This paper will involve a bit reading on publishing, on writing in archaeology, and on archaeology in the media.

I’m starting to think a bit about how this paper might connect with my interest in pseudoarchaeology. In part, I wonder whether the issues with pseudoarchaeology is less a scientific problem and more a media issue. In particular, it is interesting to think how the tactics used by pseudoarchaeologies both leverage existing communities and creating communities around their beliefs. In other words, archaeology isn’t just about “the science,” but also about “the media.”

Summer Work

I’ve started to call my summer “research leave” to help my focus on doing what I need to do and to avoid getting complacent. This summer will he hectic, in a fun way, with a few different projects rubbing shoulders with one another and it help me develop a bit of stamina for what will likely be a busy fall and winter semesters.

For those of you who wonder how the average academic spends their research leave. Here’s what I’ll be up to.

1. “Teaching as a Response to a Campus Crisis”: This paper is due August 1, but I have a substantially complete draft of the text. I think I’ll send a draft of it to a couple buddies who have endured campus budget crises in their day and see what I can do to make it stronger and more useful. I don’t have a ton of time to work on this either this summer or when I get home. I’m hoping that I can be efficient.

2. “Mobilizing the Archaeological Report for the Future Interpretive Community: Linked Open Data, Analysis, and Publication”: This is a coauthored paper with David Pettegrew for the Journal of Field Archaeology. I think we’ll work a bit on it when we’re together this summer in Greece, but most of the work on this will have to wait until September. A manuscript for review will be due September 26th, I think. So we have some time!

3. Polis I: We’ve recently learned that we need to submit the first volume of our work at Polis on Cyprus to press by the end of December (so let’s say, December 1) or risk losing funding. This is adding a much needed injection of stress to our summer work on Cyprus, but it is what it is, and fortunately, we’re close to having our part of this volume complete. In fact, most of what we need to do is the fun stuff: re-read what we’ve written and give it a bit more polish and refinement. First thing is first, though, and that’s producing a proposal for the first two volume and getting them accepted.

4. PKAP II: ARRGGGHHH… this is our long simmering second PKAP volume which is 96% done. Seriously. 96%. It is so close to being done that we could reasonably send it out for review before the end of the summer, but it has gone from being the wolf closest to the sled to just another wolf in the forest. This is less than ideal from my perspective, since I invested a good bit of energy in this volume this fall and spring, but the risk of long simmering projects is that while they might produce the richest sauce in the end, they also risk being forgotten.

5. Larnaka Sewage System pottery: This is one of those OPP (Other People’s Pottery) projects that has a spring deadline for publication. We started the work this past summer and spent some time during the “non-research leave season” collecting bibliography and strategizing how to publish this salvage material in a meaningful and efficient way. We have two weeks in Larnaka to finish our work on this material and put together some kind of very rough draft of an article to submit in the spring. 

6. Slavic Pottery from Isthmia: Last summer, we started a project to study and contextualize the Slavic pottery from Isthmia. I think our first season was moderately productive. We not only studied the material from the Roman Bath (and framed some small additional research questions), but we also came to understand both the potential and challenges of working with Isthmia data and ceramics. This summer we plan to look beyond the Roman Bath, particularly to contexts associated with the Justinianic Fortress and use these to check our contexts and typologies developed from the material from the Roman Bath. My feeling is that we’re yet another season away from producing a significant publication of this material, but we should know more or less what we want to say by the end of this summer. 

7. Hexamilion Wall Exploration Project. This is a made up name for the work that David Pettegrew and I plan to do to document what might well be some new sections of the Hexamilion Wall. We received a permit to clear some vegetation and to do some documentation and we’ll just have to see what we find. I’m optimistic. What could be very interesting is if we can connect this work with the work we’re doing with the ceramics and stratigraphy at Isthmia.

8. Publishing Work: This summer is a summer of FIVE books, I think. The Corinthian Countryside, Wild Drawing: Street Art in Perspective, The Muslims of Darürrahat, Big Pandemic on the Prairie: The Spanish Flu in North Dakota, and Clell Gannon’s Songs of the Bunchgrass Acres. I’ve never had this many irons in the fire, but I’m very excited about this bumper crop of titles scheduled to appear this fall. I’m already beginning to think of ways to market this! 

EKAS Cover-Draft 02.

9. The Slow Cooker. This fall, I’ve agreed to give a paper on my “slow cooker” idea of “Black Pseudoarchaeology.” Fortunately it is only a 10 minute paper as part of a larger workshop on Pseudoarchaeology at the ASOR annual meeting. Hopefully this gets me back to work on my next book project which will be a short book on pseudoarchaeological ideas and Black culture with particular focus on Black spiritual traditions, music, and literature. It’ll offer an alternate view to the whitewashing of the pseudoarchaeological discourse and hopefully encourage archaeologists to tread a bit more lightly when they encounter pseudo-science and pseudoarchaeological ideas in the wild. 

10. The Deep Freeze. Finally, I have a few ideas that have been shunted into the deep freeze for now. These are mostly digital projects especially related to our work at Polis. I would love, for example, to build out a digital framework and standards for publishing the archaeological data from Polis. We got a start on it may years ago so this wouldn’t be de novo. 

Three Things Thursday: On AI and LLM

At first, I was all aboard on the panic about large language models (LLM) and “artificial intelligence” in the academy. In fact, I participated in an on-campus conversation a few months ago that centered on the impact of ChatGPT in the classroom. Since then, I’ve largely grown bored of the hype and the endlessly repeated tropes that AI will change everything, we need to adapt or die, or that AI is poised to open new horizons.

Since I appreciate folks like Joshua Nundell’s efforts to respond to and critique some of the recent conversations, I thought I might add my two cents in the spirit of solidarity among bloggers, if nothing else. 

Thing the First

It seems to me that some of the anxiety surrounding the impact of LLM driven AI in the classroom centers is a bit misplaced. After all, there is a massive catalogue of approaches to writing that easily sidestep the problematic temptation to use, say, ChatGPT to produce an assignment. In my department alone, I know colleagues who do low-stakes, in-class writing, some who develop richly scaffolded writing assignments that require outlines, multiple drafts, proper citations, and other elements that LLM can’t replicate, and finally, some who encourage students to work in groups where peer pressure mitigates the risk of using ChatGPT.

Each of these approaches have pedagogical merits and are well-tested tools in a teaching tool kit. In other words, creating scenarios where LLM assisted writing is discouraged doesn’t involve re-thinking how we teach. It simply involves adopting what many have argued are “good practices” for teaching writing anyway. Of course, I understand that incorporating these practices into a class involve a bit of a redesign, but it’s hardly a revolution.

Thing the Second

Recent handwriting over the role that LLM driven AI plays in scholarship is mostly ridiculous. The examples used are, of course, egregious—especially those that preserve the telltale word, “Certainly” before listing a bunch of references in a literature review—and, presumably, embarrassing to the journals where such text appears. But let’s be honest here: these are not good journals. The examples bandied about the internet are simply not good articles as even a cursory survey of their content (and despite their being far outside my field) reveals.

In other words, this does little to convince me that a wave of AI generated content is welling up in the depths of the more unscrupulous scholarly world. Of course, most academics know that a tremendous amount of poor and mediocre scholarship exists. This is not driven by ready access to LLM derived AI composition, but by the irresistible pressures to publish frequently, to develop important quantitative markers for scholarly performance, and to constantly justify a position within the academy. Of course, publishers are only too happy to take advance of the need for content. Ironically, the pressures produced by unrealistic research expectations and unscrupulous publishers rely partly over-extended and over-worked faculty who can’t (or, more tactically, won’t) fulfill their professional obligations as reviewers.  

It seems to me that this ecosystem is as much to blame for the rise in articles that carelessly make use of LLM’s capacity to generate plausible sounding text. This isn’t to absolve the “authors” of such articles of dishonest practices, but to suggest that blaming it on ChatGPT is mistaking the symptom for the disease.

Thing the Third

Over the last dozen years, I’ve shifted from being an enthusiastic advocate for open access academic publishing to more of an agnostic skeptic. This isn’t because I think OA publishing is bad or wrong—after all I run an open access press—but that I think OA publishing as part of a more complex scholarly ecosystem that isn’t necessarily an unqualified good for all participants in this system.

It has been interesting to me to see how scholars have pivoted from championing the power of OA publications as democratizing knowledge to hesitating just a bit now as it becomes clear that OA publications may form an important component of future LLMs. Without disparaging the entire OA movement, it seems apparent that the emergence of LLM and recent challenges by copyright holders whose works constitute these LLMs creates opportunities for OA texts to create a foundation for new forms of automated and algorithmically derived knowledge making. 

Of course, for this to work, the larger ecosystem has to continue to produce high quality OA texts for our new LLM to consume. If we imagine that publishers will ultimately seek to monetize LLMs and their algorithms, then the loop is effectively closing. The growing body of OA publications, which some scholars and institutions pay to produce, will invariably populating the next generation of LLMs which will, in turn, power the next batch of AI text generators. 

This isn’t some kind of radically new observations, but does, I think, help me understand the how the larger ecosystem surrounding AI text generators and LLMs works with both teaching and publishing in the academy. 

Teaching Thursday: More on ChatGPT

Early this week, I had the pleasure of participating (and I use that word broadly) in a seminar focused on teaching with ChatGPT. The participants in the seminar were really outstanding and shared range of practical and theoretical approaches to teaching with ChatGPT. I was frankly blown away by the thoughtfulness and expertise on the panel and I struggled to engage with it entirely (despite sitting awkwardly on the “digital stage”).

Here’s what I had thought about in the lead up to the panel.

One thing that came up at the end of the panel that got me thinking a bit differently about ChatGPT is that it can only offer responses based on what it reads and consumes. On some level, this means that we have to learn to write (and disseminate) information for a new class of “artificial” readers.

This isn’t an entirely new observation. In fact, we have already understood some of this which occurs under the banner of “search engine optimization,” but it strikes me that if ChatGPT can deploy its language model to produce text, it’s not a leap to recognize that the same language model could be used to assess the content upon which its responses to queries are based. In other words, ChatGPT, and its future iterations, is a reader that, like all readers, performs its task based on an algorithm that presumably adjusts to the content available.

The question then as a writer is how do we make our work appealing or maybe merely susceptible to our new ChatBOT aggregators. Surely, these kind of bots will have less interest in deliberate displays of opacity, ambiguity, or playfulness. We might even be able to retire for good the need for the compelling (or even slightly misleading) lede which students so often turn into the cringeworthy first sentence. It also calls into question the value of such awkward stylistic crutches as the “rhetorical question.” At the same time, one wonder how it assesses the presence of “irony” in ascertaining the authority or utility of a text. A query to ChatGPT tells us that it discerns irony though linguistic features, contextual clues, semantic analysis, and, perhaps most importantly, machine learning, which relies on texts that humans have marked up to allow the AI to understand what irony looks like in practice.

On a more basic level access to texts will surely impact how these AI bots formulate their answers. For generations publishers have sought to monetize their texts by limiting access to them and recognizing that the there remains a balance between cultivating the influence of a text and capitalizing on those who need access to it. One wonder whether in the future, the scope, speed, and reach of AI readers will mean that a text that is hard to access — behind a paywall, written in a non-English language, or even is simply opaque in meaning — will limit its influence in the kind of language models that these AI readers rely on to produce new text. In other words, the presence of quick-reading AI bots will accelerate the importance of open access bodies of texts which will almost certainly gain a greater influence over the language models that shape the ability of AI bots to “think.” 

There are those who see a future where publishers are less inclined to charge for access to individual publications. Pay-wall barriers can make it harder to for automated processes to aggregate information across a wide range of sources and as aggregators increasingly serve to privilege certain sources above others (and to amplify certain works over others). Of course, there are ways to let bots in and impede the movement of human eyes, but it also stands to reason that AI-powered aggregators will invariably draw more freely on content that is more easy to access and plentiful on the web.

In light of this situation, some have proposed an alternative business model that see publishers providing aggregation services, likely powered by AI bots, that assess the significance of publications in a field, provide answers to research queries, or even prepare regular literature reviews for scholars. These aggregation services, of course, will come at a cost and will likely introduce certain biases into the results that they aggregate, but will provide a revenue stream for the publishers. More than that, as publishers shift the cost of publishing from the readers (via subscriptions) to authors (via subventions), the cost of publishing could become a way to ensure that an article rises to the top of an aggregator’s trawl through recent publications.

Teaching and ChatGPT

I know that I’m late to the bandwagon with regard to ChatGPT, but I somehow got myself featured on a panel hosted by our campus’s teaching institute to discuss what to do about and with ChatGPT in the classroom.

I think part of the reason that I’m being included is that I’m honest enough to say that I really don’t know what to do about ChatGPT, and naive enough to admit that I expect there are some great things we can do WITH it.

As for the for the former, I remain committed to some version of the “No Cop Shit” in the classroom mentality in the classroom. I’d rather have a student take advantage of my generosity than to become some kind of heavy who tries to police student efforts at resistance. Perhaps at my weakest, I believe that I’m more likely to coop a student through understanding than I am to break student resistance with force, but I like to think that my goals don’t always involve doing what I can to subvert the last of a students innate desire to resist the structures of capitalism, authoritarianism, and discipline. 

I also discovered that ChatGPT detectors are not entirely locked down. I asked ChatGPT to produce a short essay on Late Roman metallurgy on Cyprus. I then ran it through a couple of the standard AI text detectors. At least one of them told me that it was unlikely to have been produced by AI and another gave it a 50% chance of being AI generated . All the AI text detectors recognized my blog post from Tuesday as written by a human. I credit my obtuse grammatical style (once compared to Cicero on acid) and liberal sprinkling of typos. A student paper clearly composed by an AI bot, however, — the student admitted it — met with ambivalence even from ChatGPT when I asked it if its language model produced the paper. In other words, it appears that the “cop shit” route might quickly turn into a scene from Blade Runner.   

(As ChatGPT tells us “As an artificial intelligence language model, I don’t have the ability to dream or experience consciousness, so I cannot dream of anything, including electric sheep…. However, the question of what it means to be conscious and how it relates to artificial intelligence and replicants is a central theme in Blade Runner, and the movie leaves open the possibility that Deckard’s consciousness may be artificial or implanted. Overall, the nature and extent of Deckard’s consciousness are left up to interpretation and debate… there is ongoing research into the development of artificial consciousness, which aims to create machines that possess self-awareness and subjective experiences. However, such technology is still in its infancy and remains a topic of much debate and speculation… As an artificial intelligence language model, I do not possess consciousness in the same way that humans do. While I am capable of generating responses and holding conversations, these are based on algorithms and data processing rather than subjective experiences or emotions.”)

Ambiguity surrounding the character of text generated by ChatGPT especially as its language model develops (evolves?) over time makes the job of any would-be instructional blade runner at least as fraught as the administrator of the Voight-Kampf Test in the film.  

As for the potential of ChatGPT in the classroom, I remain optimistic (if a bit naive). One of the things that I’ve struggled with consistently is when a student clearly understands a topic, has done the research, and have engaged with the reading, but struggles to express their ideas in writing. Our tendency now is to work with these students to improve their writing skills, to structure their writing process, and to produce results that are adequate reflections of their ideas and engagement. This remediation comes at a cost, of course. Generally, I think it is a fair to say that students who struggle with writing, struggle academically in a college setting. It’s a hell of an environment to find yourself behind and to make up ground. In fact, in my experience students who struggle with writing often struggle academically in general because they have to invest far more time trying to write in an adequate way than students who have basic writing skills. This invariably detracts from other tasks vital to their performance in college (reading, review, problem solving, and so on). 

Of course, as the famous saying goes… all good writers are the same, but all bad writers are bad in different ways. A student who struggles with organizing their thoughts into an orthodox paper is different from a student who struggles to compose sentences despite having a well structured paper. One wonders whether ChatGPT could, in the right situation, be a crutch that allows a students whose writing in poor to avoid losing even more ground. 

A recent paper in ACS Nano, by too many authors to list, “Best Practices for Using AI When Writing Scientific Manuscripts” goes a step further and argues, as near as I can tell, that part of what makes ChatGPT convincing is that much like human generated prose, it struggles to produce the kind of bad writing that we all know (and love) from our students:

“The human-like quality of the text structure produced by ChatGPT can deceive readers into believing it is of human origin. It is now apparent, however, that the generated text might be fraught with errors, can be shallow and superficial, and can generate false journal references and inferences. More importantly, ChatGPT sometimes makes connections that are nonsensical and false.”

These problems are probably not with the language model itself, but with the text from which the language model is generated. While we may have developed beyond the idea of garbage-in, garbage-out in computing, what strikes me with ChatGPT is that it appears in my rather superficial experience with it to create text that is remarkably uniform in its badness. In other words, it produces bad text that is bad in only some, rather limited ways. In contrast, the worst student papers tend to be replete with grammatical and organization problems. ChatGPT seems to mitigate these quite effectively, but leave many of the common thinking, referencing, and evidence issues in plain sight. 

What this means for teaching is hard to know. As any faculty member who reads a considerable quantity of student work will tell you, part of the joy of reading student work is not just in its often bizarre and wonderful content, but also in its style. Students offer a window into the future of writing, thinking, and speaking English. ChatGPT seems intent on mitigating the dynamism of the English language and one wonders, at the university level at least, whether this is where it presents the greatest risk.