Every Fall, our team spends a busy few months moving between the meetings that bring the scholarly publishing community together, both virtually and in person. This season took us to ALPSP, SSP's New Directions seminar, NISO Global, the OASPA conference, the Frankfurt Book Fair, and, most recently, the Charleston Conference. Each event draws a different mix of people, yet the conversations kept returning to a shared set of questions about how research is discovered, trusted, counted, and sustained as AI becomes part of everyday scholarly work. Here are the themes that stood.

AI is now part of the infrastructure

At ALPSP, sessions on learning from industries already reshaped by AI, on AI-ready metadata, and on the pressures facing universities all pointed in the same direction. As Rebecca Hatjoullis, EMEA Regional Director for Customer Success, put it, "AI is increasingly mediating how research is discovered, and that is relevant to conversations across our client base."

By Frankfurt, that shift felt settled. Emilie Delquié, our Chief Product and Customer Success Officer, observed, “publishers have fully embraced AI and are now wrestling with how to drive revenue from it, fully seeing it as the next transition." The open questions have moved from whether to engage to how, and on what terms.

Charleston brought the library perspective into sharp focus. Kevin Forgrave, Senior Product Manager of AI, described the meeting as "three days of librarians and publishers working through what AI means for how research gets found and counted." Librarians there were clear that they want the content they license available inside the AI tools their students and faculty already use, and they want it soon.

The journal, and the judgment behind it

SSP's New Directions seminar examined how the concept of the journal would change by 2030. Program Director Nettie Lagace came away encouraged by the thoughtfulness of the panels and the positive tone in the room. "Although there will be dramatic changes in articles and journals due to AI, it was strongly felt that there is no replacement for human judgment, taste, and decisions."

Panelists acknowledged that readers' point of inquiry is moving away from the primary literature, even as the journal remains the version of record. The closing panel on business models was equally candid, as Nettie noted: "So many people are trying new things and have various pressures as the economics of publishing continue to evolve.”

Openness, trust, and provenance

At OASPA, the open access community wrestled with what openness means once machines are doing much of the reading. Sara Bowler, Senior Account Manager for Customer Success, summarized one panel's conclusion that open access still matters, but "value is moving from access to the article to provenance, trust, and machine-readable use."

The keynote sessions extended that idea. Openness on its own does not secure legitimacy; knowledge also needs to be interpretable for lay and multilingual audiences, curated against AI-generated fakes, and recognized in the ways researchers are rewarded. One moderator named curation as the thing publishers do best, a reminder that the editorial and stewardship work at the heart of publishing becomes more valuable, not less, in an AI-mediated environment. Importantly, Sara noted that "every panelist named metadata and PIDs as the plumbing that breaks everything downstream."

Counting what matters

Usage measurement was one of the liveliest threads of the fall (making our recent whitepaper on the subject particularly timely). At Charleston, a panel opened by asking who had seen usage decline over the past few years, and every hand in the room went up. COUNTER's latest guidance treats content retrieved through a publisher's own AI connection as that publisher's usage, but standardized AI usage reports are still months away, and COUNTER is encouraging libraries not to make cancellation decisions based on a temporary dip.

Hannah Heckner Swain, VP of Platforms and Partnerships, took that conversation onto the stage at Charleston as part of The Scholarly Kitchen’s “Ask the Technology Providers” panel. The panel looked at how AI agents, bot traffic, and the decline of IP-based authentication are changing both access and measurement, and it favored experimentation over resistance, noting that blocking bots too aggressively can throw away useful signals. Hannah encouraged publishers to define their risk tolerance and set measurable benchmarks, which can mean watching how AI crawlers behave before deciding how to handle them. She also concurred that publishers and libraries need new metrics that reflect agentic retrieval.

Emilie heard the same conversation from the publisher side in Frankfurt, noting that "exposing bot and LLM traffic is becoming baseline functionality." Publishers and libraries share an interest in getting this right.

Integrity and editorial capacity

As has become the theme in recent years, integrity concerns were omnipresent. At Charleston's citation integrity panel, speakers cited research on PubMed Central showing fabricated references rising from roughly 1 in 2,000 papers in 2023 to about 1 in 277 this year. Rather than rejecting those papers outright, some editorial teams are shifting toward giving authors, particularly early-career researchers, the chance to correct their references, which places even greater weight on catching problems early in the workflow.

Those pressures were also the focus of the session Emilie moderated on the Frankfurt Book Fair’s Innovation Stage, “The Pressure Point: Submission Surges, Editorial Capacity, and Infrastructure.” Drawing on our 2026 Future of Peer Review report, the conversation began with the facts: submissions reached 3.86 million in 2025, editors now desk-reject roughly two and a half manuscripts for every one they accept, and securing a single review takes an average of 4.5 invitations, nearly double the 2018 figure. The panel treated the surge as a structural challenge rooted in researcher career pressures and long-term underinvestment in editorial infrastructure, not as a technology problem waiting for a technology fix.

Looking ahead

Taken together, these meetings described an industry that has moved past the question of whether AI will reshape scholarly communication and is now working through the details of how to do it responsibly. Provenance, trustworthy metadata, meaningful usage measurement, and integrity at the point of submission are all problems that no single organization can solve alone. As Rebecca reflected after ALPSP, the real value was “the opportunity to come together in person with many customers, all approaching similar challenges together.”

We're grateful to everyone who shared their time and ideas with us this season, and we look forward to continuing these conversations with our partners and the broader community at the STM meeting next month (where you can find Dan Dyer, VP of Customer Success, and Kevin Forgrave on the stage), and on in to 2027.

1993 1999 2000s 2010 2017 calendar facebook instagram landscape linkedin news pen stats trophy twitter zapnito