This is part of my long-running commentary on the current state and future of the humanities, including what gets called digital humanities.

Nudged by a conference at Stanford

There was a symposium at Stanford last week (November 14-15) called “The Futures of Antiquity in an Age of Digital Data and AI”. Credit goes to faculty colleague Giovanna Ceserani for organizing this gathering to examine what so many are currently concerned with — the rapid rise of generative AI and the implications for the likes of the (digital) humanities.

I gate-crashed only a couple of hours — I wasn’t invited. But I was provoked enough by what I did hear, and by the program brief and paper synopses, to make this comment. What follows is not a direct critique of the symposium, but a sketch of critical matters concerning humanities and academic disciplines today.

Spoiler — there is no new digital future for the past

The title asks us to think beyond what gets called digital humanities. I welcome this. So what’s new here? My answer — certainly not a new humanities for the future.

I do not think it is useful to conceive that we are in an “age of digital data and AI”. Nor are digital data, machine learning and AI, the future of the humanities.

Questions of the digital humanities are a symptom of how institutions manage knowledge under contemporary political and economic conditions.

I list these infrastructures for building knowledge that have been rolled out over the last three decades.

I make this case and then outline the key and pressing concerns of the humanities that remain largely unexamined in the kind of discussion I witnessed last week.

Here’s the conference agenda and schedule – [Link].

The conference was framed very explicitly around a double-sided question:

• What can digital data and AI do for the study of antiquity?
• What can antiquity and its study offer to the development, critique, and governance of AI?

The program text stresses that classics has a long history with technology, but asks what genuinely new knowledge digital data and AI have enabled, especially now that AI is reshaping how we research, teach, and imagine the ancient world. It repeatedly foregrounds:

• AI as pressure on scholarly method and pedagogy (research, teaching, learning in the ancient world);

• Antiquity as a resource for imagining AI futures (eg how ancient concepts of reason, prediction, governance, friendship illuminate AI’s roles);

• A full “pipeline” of practice – from digitization of texts and images, to big-data analysis, to visualization and simulation, to heritage, ethics, and pedagogy. Panels moved from digitized literary texts, to material culture and history, to epigraphy/papyri, to heritage futures, to “AI between ancients and moderns,” and finally to philosophy and pedagogy.

The title — The Futures of Antiquity in an Age of Digital Data and AI — signals that these are big questions facing those who study antiquity — the future of the past being in question in this new (presumably) digital age. Central themes here are futures and newness.

I detect a strong emphasis on technical-methodological questions in an agenda of classic humanistic concerns: source/data quality and standards, building corpora of data and sources, bias and canon, archival selection, an interpretive hermeneutics of uncertainty, pedagogy and humanistic ethics, with newer topics of human–machine co-performance, and how to train/teach humanists in digital matters.

Classic(al) concerns, as I say.

I doubt whether such a technical focus manages to encompass the big questions implicit in the conference title — just how much is the “age of digital data and AI” affecting the humanities, given declining student recruitment, funding emphasis upon what get called STEM disciplines, culture wars around knowledge claims (eg implications of “colonial thought”), challenges to the ontology of humanities interests (eg a post-humanist critique)? Are we truly in a new age? Just how new is all this?

Let me step back and offer some answers.

Digital Humanities is not a disciplinary field — It is an institutional formation

What gets called “digital humanities” is not, and never was, a coherent intellectual field. It is best understood as an institutional formation: a convergence of funding mechanisms, administrative strategies, technical infrastructures, and labor arrangements that have reorganized how certain kinds of humanistic work are supported, branded, and made visible. The widespread claim that DH represents a methodological or epistemic rupture in the humanities does not survive even cursory historical scrutiny. What changed was not how humanists think, but how their work is infrastructured, managed, and accounted for.

The practices most often presented as distinctive of DH — corpus construction, pattern detection, classification, concordance building, mapping, modeling, editing, archiving — are not new. They are foundational to philology, archaeology, history, art history, linguistics, and anthropology. Humanists have always worked with structured data, formal models, typologies, and technologies of inscription. What digital systems introduced was scale, speed, and interoperability, not new forms of understanding. To describe this as the emergence of a new discipline is to mistake a change in material conditions for an epistemological revolution.

The mythology of DH solidified in the late 1990s and 2000s alongside major investments in digitization, tool-building, and cyberinfrastructure. New centers, labs, and grant programs — most visibly through entities such as the NEH Office of Digital Humanities — produced an administrative ecology in which DH became legible as a field. This legibility mattered: it justified funding, enabled new career paths, and provided universities with a narrative of innovation and relevance, project management focused on deliverables. But it also distorted perception. Activities that were already central to the humanities were re-described as novel once they were rendered computationally visible and administratively nameable.

This institutional consolidation brought with it a project-based logic borrowed from technoscience and grant culture: teams, deliverables, platforms, timelines, prototypes. These are not in themselves objectionable. But they have encouraged a shift in emphasis from method to technique, from conceptual work to tools, from theory to workflow. Visibility has become a proxy for intellectual advance. Dashboards, databases, and visualizations have come to stand in for arguments. In this environment, DH can present itself as the “future of the humanities” while quietly avoiding the harder work of articulating how knowledge is actually produced, interpreted, and contested.

Seen this way, digital humanities is not a disciplinary innovation but a rebranding of the humanities under new institutional, political and economic conditions. It is the humanities intensified and reorganized by digital infrastructures — nothing more, nothing less. The danger lies not in using digital tools, which humanists have always done in one form or another, but in allowing the institutional narrative of DH to obscure the deep methodological continuities of humanistic inquiry and to displace attention from the enduring theoretical questions that still define the humanities at their core.

“The Digital” and “AI” are zombie concepts

Much contemporary discussion does proceed as if we now inhabit an “age of digital data and AI,” as though this phrase names a coherent historical condition. It does not. It is a rhetorical convenience — a zombie concept that continues to circulate despite having little explanatory power, eating away at our capacity to think clearly about things. Framing the present in terms of “the digital” or “AI” exaggerates novelty, flattens history, and diverts attention from the forces that actually shape the conditions of research, teaching, and knowledge production.

Humanities scholarship has always been entangled with technologies of inscription, calculation, mediation, and automation. From writing and print to statistics, photography, film, databases, and cybernetics, there is no pre-technological humanities against which “the digital” can be contrasted. What we are witnessing today is not the arrival of something unprecedented, but the acceleration and consolidation of long-running trajectories: data construction, modeling, automation, and infrastructural dependence. To label this an “age of AI” is to mistake a phase of intensification for a civilizational break.

The appeal of these terms is not analytical but institutional. “Digital” and “AI” function as administrative signals: they attract funding, promise relevance, and align the humanities with dominant narratives of innovation emanating from the technology sector. But as concepts they do very little work. They obscure more than they reveal, especially when they are treated as causal agents — as if computation itself were driving historical change. It is not. The decisive forces shaping the humanities today are political economy, institutional restructuring, labor precarity, widening inequality, and populist challenges to reasoned discourse — conditions into which digital systems are inserted and by which they are governed.

To speak of an “age of digital data and AI” also encourages a displacement of responsibility. Structural problems facing the humanities — declining enrollments, funding priorities skewed toward STEM, managerial governance, culture-war pressures on knowledge validation — are reframed as technical challenges to be solved by better tools. This is a category mistake. No amount of machine learning will resolve the erosion of public investment in education, the making casual of academic labor, culture wars of competing conspiracies, or the instrumentalization of research under audit cultures. Technology here serves as a distraction, not a diagnosis.

This is why the concept of “the digital” has become something of a zombie: endlessly invoked, rarely interrogated, and incapable of accounting for the realities it is supposed to explain. The concept survives because it is useful to institutions, not because it clarifies thought. By foregrounding technology, we background the long-standing questions that actually matter: how knowledge is made, who it serves, under what conditions, and toward what futures. The humanities do not need to be saved by AI. They need to confront, once again, the enduring problem of how to think, interpret, critique, and act within a world structured by power, inequality, and historical inheritance.

What Digital Humanities often avoids: theory, method, and the big questions

One of the most striking features of what passes for digital humanities is not what it addresses, but what it systematically avoids. Much DH discourse is preoccupied with tools, workflows, datasets, and technical capabilities, while leaving largely unexamined the deeper questions of theory and method that have always anchored humanistic inquiry. The result is a field that is busy, productive, and often technically impressive — but conceptually thin.

At the core of the humanities are long-standing questions about how knowledge is made and justified: interpretation and explanation, understanding and modeling, description and critique. These questions are not rendered obsolete by computation. On the contrary, they become more urgent as scale, automation, and abstraction increase. Yet much DH work proceeds as if methodological reflection were optional, or worse, as if it had already been settled by the availability of new techniques. Pattern-finding is treated as insight. Visualization is taken to be argument. Scale is mistaken for understanding.

This avoidance is particularly evident in the treatment of AI. Large language models, clustering algorithms, and image-recognition systems are often presented as if they were interpretive agents, rather than statistical machines that reorganize existing material. They retrieve, correlate, summarize, and simulate — but they do not explain, understand, or judge. Those acts remain human responsibilities, grounded in historically situated reasoning. When DH fails to articulate this distinction clearly, it risks reviving old positivisms under the guise of technical sophistication.

Equally absent is sustained engagement with social theory, political economy, and the arts as modes of knowledge. The humanities have never been isolated from the social sciences nor from creative practice, nor from the ontological and epistemological concerns of the hard sciences. Archaeology, in particular, has always occupied this hybrid space, working simultaneously with material evidence, models of social process, narrative interpretation, and speculative reconstruction. Yet DH discourse often reinstates disciplinary silos by focusing narrowly on technique and leaving untouched the conceptual frameworks that allow us to connect material and cultural forms, social structures, and historical change.

What is lost in this narrowing is precisely the humanities’ capacity to address big questions: what counts as evidence; how pasts are made present; how models mediate reality; how narratives shape understanding; how power operates through archives, classifications, and infrastructures. And the biggest of them all — just how are we to conceive of the human in a world of things, of other species, of non-humans? These are not ancillary concerns. They are the work of the humanities. When DH sidelines them, it becomes not an advance but a retreat — away from method, away from theory, and away from the intellectual responsibility to explain why any of this matters.

In short, digital humanities too often treats technique as a substitute for thought. Concepts follow tools, if they appear at all. But method does not emerge from software, and theory cannot be automated. Without sustained attention to these foundations, DH risks becoming an elaborate technical service layer attached to the humanities rather than a serious engagement with their enduring intellectual challenges.

A counter-genealogy: reflections on personal experience

The symposium nudged me to reflect upon my own experiences in matters disciplinary and digital. Here are a few words of orientation — a reminder to myself, as much as anything else, of a personal itinerary from classical studies to a contemporary archaeological sensibility.

If the dominant story of digital humanities is one of rupture and novelty, my own career traces a different genealogy — one of continuity rather than disruption, of method rather than technology, of practice rather than branding. Over more than four decades, my work has repeatedly intersected with new media, computational tools, and digital platforms. But at no point did these encounters pull me into a new field called “digital humanities.” They simply offered new materials and new conditions for doing what the humanities — and archaeology in particular — have always done.

After a very traditional grammar-schooling in classical languages I shifted direction, at the protest of my teachers, into the archaeology of classical antiquity. A bigger picture was what I was seeking, and one that reached beyond the elite culture-industry of then-and-now.

I am currently working with Gabriella Giannachi on the intersection of arts practice and AI, part of an extensive government-funded research network in the UK [Link]. In between I have explored GIS, VR and AR in learning, digitally-facilitated research collaboration (Web 2.0), augmented reality archives, hybrid analogue and digital media ecologies. In the last couple of decades most of this has been through my studio at Stanford and when I fronted Stanford Humanities Lab with Jeffrey Schnapp and Henry Lowood (up to 2009). Speculative fabulation. Theatre archaeology. Applied humanities. Design foresight and futures literacy, from Rotterdam to Aisin corporation. Scholartistry and Creative Pragmatics — active and hybrid learning beyond a sciences-humanities split — a series of projects with Connie Svabo and her lab at University of Southern Denmark [Link].

In the late 70s I explored the potential of statistical analysis of archaeological data associated with early farming communities in northern Europe — affirming patterns in large datasets of prehistoric mortuary practices: distributions of grave goods, spatial arrangements, demographic associations, and formal regularities. This was a computational humanities only made possible by the availability of high-end main-frame CPUs that could cope with the number of calculations involved in multi-variate analysis. The aim was never to replace interpretation with numbers. Statistics were lenses — ways of sharpening attention to structure and variation in material traces. Pattern-finding required judgment, contextual knowledge, and, above all, theoretical framing (structural-marxist socio-cultural modeling. Computation changed scale and speed, not the logic of inquiry.

The same is true of my later work with GIS. Geographic information systems did not introduce a new epistemology; they intensified an old one. Mapping is a hermeneutic and eidetic (pragmatic) practice. It directs attention, foregrounds relations, and encodes assumptions about space, agency, and causality. GIS taught me something elemental that has stayed with me ever since: digital tools reshape perception, attention, and workflow, but they do not think for us. They help us see differently, not understand differently.

This conviction shaped my approach when I co-directed Stanford Humanities Lab. We were explicit — and insistent — that SHL was not a center for “digital humanities.” It was a laboratory in the arts and humanities, embedded where appropriate in the social and physical sciences, and conceived as a studio for experimental scholarly practice. Our projects were not DH projects. They were humanistic projects that happened to use digital media: multimedia archives, collaborative platforms, performance-based interpretation, experimental reconstructions, and interactive models. We did not claim new methods. We made visible the humanities’ long-standing multimodality.

The same logic guided my work on collaborative platforms such as Traumwerk in the early days of Web 2.0. Before “platforms” became a managerial buzzword, this work explored distributed authorship, open commentary, and dialogic scholarship. Again, there was nothing fundamentally new here. Medieval glossators, early modern republics of letters, and nineteenth-century philologists all worked in commentarial traditions. Digital systems simply redistributed those practices across time and space.

My subsequent work in theatre/archaeology, archaeography, and arts practice as research pushed these commitments further. Performance, scenography, model-making, and speculative reconstruction are not embellishments to scholarship; they are modes of inquiry in their own right. They make worlds in order to think with them. In these contexts, method appears not as a formal protocol but as design—as the orchestration of concepts, materials, narratives, and audiences in situated projects. This concern with design connected with archaeology conceived as material-culture studies, and took me into design foresight, complementing direct collaboration with designers and artists.

Most recently, my engagement with AI and in another project with Gabriella Giannachi, has taken the form of what I call “archaeologies of AI”: treating intelligence, automation, and machine learning as cultural and historical phenomena, embedded in archives, infrastructures, and political economies. Once again, the question is not what the technology can do for the humanities, but how humanistic methods —interpretation, critique, genealogy, imagination — can clarify what these systems are, how they work, and what futures they help bring into being.

Taken together, this counter-genealogy shows something simple but often forgotten: the digital is a material condition of contemporary humanistic practice. New tools intensify capacities that were already there. They do not displace the need for theory, judgment, or responsibility. If anything, they make those needs more pressing.

Creative Pragmatics: method as practice, not paradigm

What this counter-genealogy points toward is not a rejection of abstract theory, nor a retreat into empiricism, but a different understanding of method itself. Connie Svabo and I have come to call this orientation Creative Pragmatics [Link]. It treats theory not as adherence to paradigms or schools of thought, but as a repertoire of conceptual tools deployed within situated projects. Method, on this view, is not abstract prescription. It is practice — designed, enacted, revised, and learned in the course of doing work under real constraints. In its transdisciplinary practice Creative Pragmatics transcends the regular distinction between arts, humanities, social sciences, physical sciences.

Creative Pragmatics draws on science and technology studies, design practice, archaeology, performance, and rhetoric. It starts from a simple observation: research does not unfold in the abstract. It happens in institutions, with particular resources, skills, audiences, technologies, and temporal horizons. Concepts are not detached explanations; they are tactics — ways of directing attention, framing questions, staging evidence, and making sense of what remains in the performance of knowledge. Theory is inseparable from mediation, from inscription, from the forms in which knowledge is made public and consequential.

Seen this way, the familiar opposition between humanities and social sciences, for example, collapses. In practice, both work through pattern-finding, modeling, interpretation, narrative construction, and critique. Both rely on judgment and imagination as much as on formal procedure. Archaeology has long exemplified this hybridity: it is simultaneously empirical and speculative, analytical and creative, explanatory and interpretive. Digital tools do not change this. They intensify it.

This is why the most important methodological questions raised by AI and large-scale computation are not technical. They concern praxis, thoughtful practice: how humans work with machines; how uncertainty, ambiguity, and interpretation are handled; how models are situated within arguments; how responsibility is distributed across human and non-human actors. Learning to use AI critically is not a matter of mastering software. It is a matter of cultivating judgment, reflexivity, and ethical awareness in environments saturated with automation.

Creative Pragmatics therefore places learning and pedagogy at the center. Research is inseparable from learning by doing. Knowledge is produced through projects, experiments, performances, and interventions, not simply through the application of tools. Futures literacy [Link], in this sense, is not about keeping up with the latest technologies. It is about developing the capacity to orient oneself within changing infrastructures, data regimes, and media ecologies—to ask, repeatedly and critically, what kinds of pasts and futures are being made possible.

If digital humanities has a future worth claiming, it lies here: not as a field defined by technology, but as a space in which the humanities recover and renew their methodological seriousness. Concept comes before technique. Practice comes before branding. And method is understood, once again, as a creative, pragmatic, and fundamentally human undertaking.

What is the past to become?

The humanities are not in crisis because they have failed to adopt digital tools. They are in trouble because institutions have struggled to articulate what the humanities are for in a world shaped by accelerating inequality, extractive political economies, environmental breakdown, challenges to reasoned argument, and contested futures. Framing the problem as one of technology — of catching up with AI, data, or “the digital” — is a way of avoiding this harder reckoning.

Digital humanities does not fail because it is insufficiently technical. It fails because it was too easily institutionalized without a corresponding deepening of methodological and theoretical ambition. In many cases it substitutes infrastructure for vision, tools for concepts, and visibility for understanding. That does not make DH useless. It makes it incomplete — and in need of re-situating within a broader account of how humanistic knowledge is made, taught, and mobilized.

The question that matters now is not whether AI will transform the humanities. Of course it will, in banal and uneven ways, as all media do. The real question is older and more demanding: what is the past to become? How are histories, archives, and cultural inheritances mobilized in the present, and toward what futures? Who controls the infrastructures through which knowledge circulates? What kinds of judgment, imagination, and responsibility do we want to cultivate in students and scholars working amid automation?

Answering these questions requires intellectual leadership, not technical enthusiasm. It requires treating method as something to be argued for, practiced, and taught — not assumed to emerge automatically from new tools. It requires recognizing that the humanities have always been interdisciplinary, experimental, and engaged with technology, but never reducible to it.

If there is a future worth defending for the humanities, it lies in reclaiming this seriousness of purpose. Not in declaring new ages. Not in chasing zombie concepts. But in doing what the humanities have always done at their best: working critically and creatively with the remains of the past in order to imagine better, more just, and more thoughtful futures.

Animating the archive 2005. Life-squared — replaying Lynn Hershman-Leeson’s 1972 installation at the Dante Hotel San Francisco in the online world Second Life. Stanford Humanities Lab for the Presence Project (Arts and Humanities Research Council UK 2005–2009).