This is a recap of the Prophy Predicts webinar—"The Future of Scholarly Content"—hosted on May 28, 2026. The full recording is available on YouTube: https://youtu.be/l5I_yN3a8cY.
For more than 350 years, the journal article has been the primary vehicle for communicating research. It survived the move from manuscript to print and from print to digital. Now it faces a new challenge: AI systems that increasingly mediate how people discover, consume, and use scholarly information.
In episode four of Prophy Predicts, we brought together four practitioners who work inside the scholarly publishing system to ask where it is heading. They were asked the same questions in advance: do you still believe in journals? What about books? Where do you see scholarly content in two to five years?
Their answers were very different—and that diversity of opinion revealed just how unsettled the future of scholarly publishing remains.
Prophy Predicts is a monthly series of live panel discussions on the future of scholarly publishing. Episode four featured:
Moderated by Gareth Dyke, Partnership Director at Prophy.
Ashutosh Ghildiyal opened by reframing the question. Rather than asking whether journals will survive, he asked what they are actually for—and worked through six purposes that scholarly content has historically served: registration, verification, availability, discoverability, dissemination, and archiving.
His argument: journals will continue to fulfil the most critical of these—maintaining a verified and verifiable record of research—but the other functions are increasingly being taken over by other parts of the ecosystem. AI, specifically, is becoming the primary interface through which researchers access and apply information.
"AI systems will become the primary means of consumption of research," he said. The implication is significant: journals may no longer be the primary discovery layer. Instead, they become the trusted corpus that AI systems draw on, interpret, and synthesise.
In this model, publishers become providers of trusted context: verified interpretation, plain-language summaries, and structured knowledge that AI systems can use more reliably.
Jennifer Mahar has been a managing editor for over 25 years. She still believes in journals, and she was clear about why: peer review has fundamental value, and that isn't going away.
But her prediction focused less on preservation and more on acceleration. She described an editorial landscape under real strain—manuscripts arriving faster than existing workflows can process them—and argued that automation is not optional but necessary. Trust markers, in her framing, are practical workflow tools: levers that handle the mechanical checks so human editorial effort can be redirected to the work that actually requires human judgment.
She pointed to a session she had attended at SSP that same morning, where automated checks for PRISMA and CONSORT compliance were being demonstrated. And at the Society of Surgical Oncology, she described a practice already in place: every paper published in Annals of Surgical Oncology includes a visual abstract. "We're really trying to make sure that we're communicating the information in many different ways."
Her most pointed prediction was about format. Journals, she argued, need to move into short-form multimedia at a pace the industry has not yet matched. "These reels and shorts and ways of communication are going to be very important as we move forward."
Katie Corker offered the most structurally critical perspective on the panel. She leads ASAPbio, a nonprofit focused on faster and more open research communication, primarily in the life sciences.
Her position is that peer review remains essential, but where and how it happens needs to change. The model she described is the publish-review-curate approach: research is shared first on preprint servers and open repositories, then open evaluation is layered on top. "It's not abandoning peer review, but it is changing where peer review occurs in the workflow."
She also pushed back on how trust markers are currently being discussed. Rather than branded indicators that signal "you can trust this" without further detail, she argued for transparency: trust markers should be fine-grained and open, so that readers can see exactly which checks were performed and draw their own conclusions about what that implies.
Her closing remark captured the orientation she brought throughout: "The current system is not sort of neutral or permanent, and it's very, very expensive."
Lyndsey Dixon joined from Hong Kong, advising publishers on M&A and strategy. Her case for journals was built on three points: community, infrastructure, and trust.
On community: she argued that journals are communities first and publications second. Editors, reviewers, authors, and editorial board members are often the same people at different moments in the same conversation. "Publishing is people." That human infrastructure, she argued, will matter as much as the content itself as the field changes.
On infrastructure: she pointed to Henry Oldenburg's founding of the Philosophical Transactions of the Royal Society in 1665, which specified four functions—registration, verification, dissemination, and archiving. Her contention: no single entity currently replaces all four. Preprints address some; none address all.
On trust: she made the argument that scarcity of credibility increases its value. The Michelin Guide does not become less relevant because there are more food blogs—it becomes more relevant, unless it starts approving mediocre restaurants to chase volume. "The ones that will really matter are the ones that double down on trust, accountability and community."
She was more equivocal on books. Her prediction was that the monograph market is bifurcating: prestige long-form humanities scholarship on one end, and everything else migrating to other formats. "That solid-but-unremarkable academic monograph is probably the format most at risk." Most sell under 300 copies. Discovery is broken at that end of the market.
All four panelists saw trust becoming more valuable as AI-generated and low-quality content proliferates. As the volume of research grows, the value of a credible quality signal increases—but only if the people issuing that signal are willing to maintain it.
The second convergence point was on the role of societies. All four identified scientific societies—not journals per se—as the locus of genuine research community. The society convenes the field; the journal is one vehicle through which it communicates.
The third was on AI training. Both Ashutosh Ghildiyal and Lyndsey Dixon flagged the feedback loop problem: AI trained on low-quality or unverified content degrades. Publishers who supply AI systems with clean, structured, verified scholarly knowledge make themselves indispensable to the architecture, rather than being displaced by it.
On workflow: Automation of routine checks is no longer a future investment—it is operational. The question is how to redeploy the human capacity that automation frees.
On format: Short-form multimedia and modular content complement the journal article by helping research reach audiences that may never read the full paper.
On AI: Publishers who position themselves as trusted context providers—supplying AI systems with verified, structured knowledge—have a different business model than publishers who compete with AI for the same users.
On trust: The credibility of the evaluation process may become one of the most valuable assets publishers can offer.
Prophy Predicts is a monthly live panel series on the future of scholarly publishing. Each episode brings together practitioners from across the industry to share predictions and challenge one another's assumptions.
Watch the full recording: https://youtu.be/l5I_yN3a8cY
Prophy helps publishers identify relevant reviewers, detect conflicts of interest, and discover related research using a database of 98M+ researcher profiles and 192M+ publication records.
Learn more at prophy.ai