I was in the room at the Frankfurt STM meeting in 2015 when my friend and now colleague Roger Schonfeld delivered a talk that changed how I understood the problem I thought I had been working on for most of my career.

Roger mapped the researcher scholarly journal access experience in plain terms. Journals were online. DOIs worked. Link resolvers existed. The digitization problem had been substantially solved. But the access problem had not. Researchers with legitimate institutional access were still bouncing between discovery tools, publisher platforms, authentication systems, and paywall dead ends for content their institutions had already paid for. Many were finding easier paths through ResearchGate, preprint servers, and other informal routes, often without realizing that convenience came at the cost of authority, currency, or connection to the Version of Record.

Roger’s conclusion was uncomfortable: this was not a series of isolated product gaps. It was a systemic failure. No single publisher, platform, library, or vendor could fix the seams between discovery, authentication, entitlement, versioning, and access. The fix would have to be collective. At that same meeting, a group of us publishers huddled around that problem, and over time that conversation evolved into GetFTR (Get Full-Text Research)—a practical attempt to address those access pathway issues through shared infrastructure.

That talk landed with me because I had spent years working at exactly that point of friction. I joined Wiley in 2007 as the company’s first director of user experience, with the explicit job of understanding how researchers experienced scholarly publishing and fixing what was broken. Roger helped me see that the work I had been doing—making content better and more accessible within a publisher’s own systems—was necessary, but not sufficient. The deeper problem lived in the seams between platforms, institutions, incentives, and standards.

In 2026, AI has brought that lesson back with new urgency.

AI systems are beginning to mediate scholarly workflows: discovery, summarization, evidence synthesis, peer-review support, research assessment, learning, and institutional decision-making. But they are entering an ecosystem built for human navigation, not machine action. They need to know which version of a work they are using. They need to know whether it has been corrected, retracted, peer reviewed, superseded, licensed, or restricted. They need structured ways to retrieve authoritative content, respect rights, preserve provenance, and carry integrity signals into downstream workflows.

Too much of that information remains fragmented, inconsistent, platform-specific, trapped in PDFs, or absent from machine-actionable infrastructure altogether. That is not simply an AI problem. It is a scholarly communication problem.

And once again, the fix will have to be collective.

This is why I joined Ithaka S+R.

My role is to help build a new practice focused on AI in scholarly communication: helping publishers, societies, infrastructure providers, platform vendors, and other scholarly communication organizations develop AI strategies and products that are not merely powerful, but scholar-ready. By that I mean AI systems that reflect integrity, transparency, rights awareness, provenance, and institutional trust.

That phrase, scholar-ready AI, matters to me because it shifts the question from what AI can do to what scholarship requires.

A tool may summarize quickly and still be unusable for serious research if it cannot distinguish a preprint from the Version of Record. A discovery agent may retrieve relevant-looking evidence and still be unsafe if it cannot carry retraction, correction, or peer-review status into the answer. A publisher may build an impressive AI product and still struggle in the market if libraries and universities do not have clear ways to evaluate, trust, and procure it. A library may want to support AI-enabled research workflows and still lack shared methods for determining which tools meet scholarly expectations.

These are not only product questions. They are strategy, policy, infrastructure, metadata, rights, integrity, procurement, and governance questions. They sit exactly at the boundary between what publishers and tool providers can supply and what libraries, researchers, universities, and institutions need to trust.

That bridge between supply and demand is what makes Ithaka S+R the right home for this work.

I am joining with nearly two decades of experience on the publisher side: building platforms, leading product strategy, working through rights, metadata, access, AI, and infrastructure questions from inside a major STM publisher. But AI strategy in scholarly communication cannot be solved from the supply side alone. Publisher and vendor offerings will only succeed if libraries, researchers, universities, and other institutional partners can understand them, evaluate them, trust them, and procure them.

At Ithaka S+R, I will be able to connect what I know from the publisher side with my colleagues’ deep engagement with the demand side of the ecosystem—the institutions and communities whose expectations should shape what scholar-ready AI becomes.

That is the work I came here to do: to help bring those two sides into a more productive conversation. Publishers, societies, platforms, and infrastructure providers need practical guidance on AI strategy, product architecture, rights, metadata, provenance, and integrity. Libraries, researchers, and universities need ways to evaluate which AI tools are trustworthy, evidence-based, rights-aware, and aligned with scholarly norms. Ithaka S+R sits in the rare position of being able to bring those questions together.

The work ahead will be practical. It will include advisory services, frameworks, evaluation methods, rubrics, scorecards, playbooks, and convenings that help organizations make better decisions about AI strategy and implementation. It will mean helping the supply side understand institutional expectations. It will mean helping the demand side articulate what trustworthy scholarly AI should require. And it will mean working across the field to turn those expectations into shared methods, not just shared anxieties.

ITHAKA has been doing this kind of field-level work for thirty years. JSTOR began with a problem no single library could solve alone: the physical, economic, and preservation burden of scholarly journals at scale. A neutral organization helped transform that shared problem into durable infrastructure. Ithaka S+R has continued that tradition through research and advisory work that helps libraries, publishers, universities, and other organizations understand and navigate change before it hardens into crisis.

That history matters now. The AI transition does not need another vendor pitch, another platform roadmap, or another abstract principles statement. It needs practical, trusted, cross-sector work grounded in how scholarship actually happens and how institutions actually make decisions.

When people asked me where I was going after Wiley, I found myself saying it was something like a think tank. That answer was not quite right.

Ithaka S+R is the place where the problems I have been circling for a decade are not adjacent to the work. They are the work.

I can’t wait to get started.