Unlocking the Future of Scholarly Publishing: How Generative AI Is Redefining Value and Revenue

Feb

Scholarly publishing is at an inflection point.

For decades, STEM society publishers have depended on journals, conferences, and membership as their economic backbone. Today, generative AI and large language models (LLMs) are accelerating a structural shift in how knowledge is created, discovered, and monetized. The central question is no longer if AI belongs in the publishing value chain, but rather how we responsibly deploy it to advance our missions and diversify revenue.

The AI Imperative: Why Scholarly Publishing Can’t Afford to Wait

AI is rapidly transforming the economics of content.

As Darla Henderson (FASEB) emphasized at the ACCESSE25 meeting, “the genie is not going back in the bottle.” STEM societies need to move beyond fear and uncertainty, openly address concerns, and connect generative AI to real pain points in publishing. Waiting for a “perfect” AI moment risks relevance erosion, as commercial platforms race to build knowledge experiences that sit between researchers and society content.

During ACCESSE25, Rob Barnes (Betty:AI) also underscored that associations are uniquely positioned for this moment: “STEM societies have the largest repositories of knowledge for any industry or profession. The systems and processes that these associations build around creating knowledge (peer review, collaboration, human-in-the-loop) will be incredibly important, and never more so than now with generative AI.”

How to Start Building the Foundation: Data, Governance, and Trust

Data Readiness: During the panel with Darla and Rob, Chad Stewart (MemberJunction) stressed that “data is the fuel for AI.” Societies must unify and clean their data (structured and unstructured) to unlock new insights and applications. This means harmonizing metadata, consolidating repositories, and tagging assets for discoverability. “Garbage in, garbage out” is now “hallucinations in, hallucinations out.” Bottom line: Quality data is essential for reliable AI outputs.

Governance: Establish clear policies for privacy, security, and intellectual property. As Darla shares, this requires a working group with both advocates and skeptics, broad stakeholder engagement, and consensus-driven policy development. Societies should encourage experimentation, balanced with security and mission alignment.

Trust: Rob Barnes also highlighted that “the trust that has been established over 100 years or so in a body of knowledge… it’s never been more important for an AI to be an extension of that trust.” Specialized AI agents trained only on a society’s content can protect IP and maintain quality, while human-in-the-loop review and transparent metrics reinforce credibility.

Monetizing AI: From Licensing Models to New Services

Roy Kaufman (Copyright Clearance Center) makes this point clear: “AI is changing the market landscape. Materials, copyrighted materials, are being made, copies are being made and stored. But if you’re not licensing, your content is in play.” Societies must proactively engage in licensing to protect their assets and generate revenue.

Revenue Paths for STEM Societies to Explore Include:

Responsible Data Licensing: Annual licenses and recurring revenue models are possible when societies control their rights and negotiate terms that recognize ongoing content creation.

Premium Knowledge Interfaces: AI-powered conversational search, synthesis-on-demand, and personalized reading lists can sit behind member or institutional tiers, increasing the value of membership and reducing “time to insight.”

AI-Assisted Editorial Services: Tools for terminology checks, reference validation, and reporting guideline assessments streamline workflows and can be monetized as new services.

API Access: As Rob Barnes described, “You’re selling access to the API (e.g., Betty’s use of your knowledge to answer questions), not just the knowledge itself.”

Real World Example: The American Geophysical Union (AGU) recently partnered with Wiley and a European Space Agency AI project to create a virtual expert that answers questions and cites AGU content, demonstrating a refined, citation-aware use of society knowledge.

Lessons from the Frontlines

Start Narrow and High-Value: Focus on a bounded use case (e.g., conversational discovery for one flagship journal) and define clear success metrics.

Build with the Community: Include editors, authors, reviewers, and members in co-designing interfaces and guardrails.

Measure Mission Outcomes: Track how AI features reduce researcher effort, speed editorial decisions, and increase cross-content discovery.

The Road Ahead: Preparing for an AI‑First Future

AI adoption should be treated as a program, not a project. A recommended roadmap for STEM societies includes:

Data readiness, policy definitions, and a tightly scoped pilot with human review.

Integration with identity and access systems, rollout to priority member segments, and monetization experiments.

Portfolio expansion, continuous model evaluation, and ongoing governance updates.

Partnerships also matter, so work with vendors who are transparent about model sources, safety layers, and IP posture. And above all, keep your members informed.

Lead With Mission, Measure With Outcomes

Overall, generative AI is less a shortcut and more a new operating system for how societies organize, interpret, and deliver knowledge.

The winning strategy is clear: anchor in trust, prepare your data and governance, pilot high-value use cases, and monetize through responsible licensing and member-centered services.

STEM societies that move now will strengthen their role in the research ecosystem and create durable, diversified revenue without compromising editorial integrity.

An approachable next step may include exploring what single workflow—research discovery, standards interpretation, or editorial support—would most improve your members’ daily work. Start there, then measure rigorously and share what you learn with your team.

Acknowledgement

This piece was informed by the content presented during the ACCESSE25 session “Generating New Publishing Revenue in an AI Powered World,” led by:

Rob Barnes, CAE, AAiP AUPG, CEO and Co-Founder, Betty:AI

Darla P. Henderson, PhD, Executive Director and CEO, Federation of American Societies for Experimental Biology (FASEB)
Roy Kaufman, Managing Director, Business Development and Government Relations, Copyright Clearence Center