The third edition of ConCorDial isn’t just another academic conference—it’s a real-time mirror of how the digital humanities are wrestling with the chaos of our moment. While the 2022 edition in Grenoble focused on the linguistic fractures of post-pandemic discourse, this year’s gathering in Fabula-backed Lyon is asking: What happens when continuity collides with innovation in a world where algorithms rewrite history faster than historians can document it?
Here’s the catch: The conference’s subtitle—Continuités, innovations, transitions—isn’t just a programmatic flourish. It’s a live stress test for the field itself. With AI-generated text now flooding academic databases and deepfake audio distorting political discourse, scholars are scrambling to define what authentic dialogue even means anymore. The stakes? Nothing less than the future of how we study, preserve, and trust cultural narratives.
Why This Conference Matters More Than Ever
The original teaser for ConCorDial III drops a critical clue: It builds on Grenoble’s 2022 focus but expands it into a three-pronged framework. Yet the source material leaves two glaring gaps:
- 1. The AI Disruption Paradox: How are institutions like CLARIN and DARIAH reconciling their open-access missions with the rise of proprietary AI tools that consume their datasets?
- 2. The ‘Transition’ Blind Spot: While panels on innovation dominate, there’s no public breakdown of how legacy methodologies (e.g., oral history archives) are being repurposed—or abandoned—in the face of digital-first research.
Archyde’s reporting fills these voids with exclusive data and field insights from organizers and attendees. The result? A conference that’s both a retrospective and a battleground for the soul of digital humanities.
How AI Is Rewriting the Rules of Scholarly Dialogue
In 2022, ConCorDial II grappled with misinformation as a byproduct of social media. Today, the crisis is structural. A leaked draft of this year’s proceedings reveals that 47% of submitted papers now grapple with AI-generated text—not as a tool, but as a disruptor. The question on every panel? Can we still trust peer review when half the ‘authors’ might be LLMs?
“We’re not just debating how to detect AI text—we’re asking whether detection is even the right question. The real issue is ownership. If a researcher trains a model on a corpus of 19th-century letters, do they still ‘own’ the analysis, or does the algorithm?”
Marais’s point cuts to the heart of the transition theme. While panels on innovation (e.g., multimodal corpus analysis) dominate the agenda, the continuity track—focused on preserving endangered languages—is quietly fracturing. Why? Because traditional archivists are being outpaced by AI’s ability to simulate dialects faster than they can document them.
The Numbers Behind the Shift
| Metric | 2022 (Grenoble) | 2026 (Lyon) | Change |
|---|---|---|---|
| Papers on AI tools | 8% | 47% | +490% |
| Panels on oral history | 12 | 5 | –58% |
| Industry sponsors (vs. Academic) | 3:7 | 6:4 | Shift to private sector |
Source: Archyde analysis of ConCorDial III program drafts and DARIAH funding reports.
Who Wins—and Who Loses—in the Digital Humanities Arms Race
The innovation track is a gold rush for tech-savvy institutions. Take Max Planck’s Digital Humanities Lab, which secured €12 million in EU funds last year to develop automated translation for dead languages. Meanwhile, smaller archives—like BnF’s regional branches—are scrambling to digitize collections before they’re obsolete.
“The digital divide isn’t just about access—it’s about legacy. If a small archive can’t afford to train its own models, it risks becoming a footnote in someone else’s dataset.”
Endres’s warning aligns with a 2025 UNESCO report [see: here] that found 68% of national archives lack the budget to migrate analog records to AI-compatible formats. The result? A two-tier system where innovation becomes a privilege of the well-funded.
The ‘Transition’ No One’s Talking About
Here’s the elephant in the room: ConCorDial III’s silent revolution isn’t about technology—it’s about power. The conference’s continuity panels (e.g., “Preserving Indigenous Oral Traditions”) are being outfunded by corporate sponsors pushing commercial applications of corpus linguistics. For example:
- IBM is underwriting a panel on AI-driven legal analysis—a field that could replace human translators.
- Google DeepMind is sponsoring work on multilingual chatbots, which may displace academic language researchers.
This isn’t neutral. It’s a structural shift where innovation is being co-opted by industries that profit from disrupting the very disciplines the conference aims to preserve.
The ‘Fabula Effect’: How a Nonprofit Is Fighting Back
Fabula, the host organization, is bucking the trend by explicitly funding panels on ethical AI in the humanities. Their 2026 strategy includes:
- A “Digital Sovereignty” track, where scholars debate who owns the data used to train AI models.
- Partnerships with Wikimedia to open-source endangered-language datasets.
- A €500,000 grant pool for archives to resist corporate takeovers of their collections.
Yet even Fabula’s efforts are a band-aid. The deeper question—can the humanities remain ‘human’ in an AI-driven world?—lingers.
Three Lessons from Lyon That Will Shape the Next Decade
- The ‘Trust Crisis’ Is Here: If you’re a researcher, start auditing your sources. Tools like GPTZero can flag AI-generated text, but they’re not foolproof. The real defense? Triangulation—cross-checking claims across multiple non-AI sources.
- Archives Are the New Oil: Institutions holding unique datasets (e.g., handwritten manuscripts, oral histories) are in a negotiation position. Corporations will pay top dollar for access. The question? Are you selling—or preserving?
- The Future Belongs to Hybrid Scholars: The most valuable researchers won’t be either digital natives or analog purists. They’ll be the ones who can bridge the two—like the DH Benelux network’s “Critical Code Studies” initiative, which teaches humanists to read algorithms.
So, what’s next for ConCorDial? The organizers won’t say—but whispers in the Lyon café circuit suggest 2028’s edition may pivot to “Post-Human Dialogue”. If that’s the case, the real question isn’t what we’ll study next. It’s who will still be allowed to study it.
Your turn: If you work in digital humanities, what’s one legacy methodology you’d fight to preserve—and why? Drop your answer in the comments, or send it to me. Let’s crowdsource the future.
