Last week, I attended the 47th annual meeting of the Society for Scholarly Publishing (SSP) in Baltimore. As always, it was a great opportunity to reconnect with familiar faces, meet new voices in the field, and get a sense of where scholarly publishing is heading. One thing became abundantly clear across the three days: AI isn’t just the hot topic of the moment — it’s reshaping our industry, fast.

Compared to just a few years ago, the mood around AI has shifted dramatically. Back then, generative AI tools like ChatGPT were seen by many as a threat — to authorship, to peer review, and even to the credibility of scholarly content itself. Fast forward to 2025, and the tone has changed. AI is no longer a curiosity or a cautionary tale. It’s here to stay — and it’s increasingly being integrated into editorial and publishing workflows in thoughtful, ethical ways.

From skepticism to smart implementation

A standout session at SSP this year was “Navigating the AI Frontier: Developing Robust Governance Policies for Publishers, Authors, and Reviewers.” In it, speakers from the American Institute of Physics and the American College of Physicians shared concrete examples of how AI is already supporting their processes. From automating metadata enrichment to enhancing reviewer suggestions and even helping assess submission quality before peer review, these organizations are showing what’s possible when AI is approached with intention and accountability.

The message was echoed in the session “What Smaller Publishers Need from Tech Vendors to Level Up.” Smaller organizations — often without dedicated AI labs or in-house data science teams — are looking to vendors for tools that are not just powerful, but trustworthy and usable. This is where the right foundation matters.

Trust, transparency, and structure

There’s an important common thread here: successful, responsible AI in scholarly publishing depends on structure. Not just in governance and policies, but in content itself.

Structured content — especially XML-based, semantically rich content — provides the foundation that AI systems need to work effectively and ethically. Why? Because AI, at its core, is only as good as the data it learns from and the context it can understand. If content is ambiguous, unstructured, or locked in PDFs and Word files, AI systems struggle to generate meaningful or trustworthy output. Worse, they may introduce hallucinations or misinterpretations.

But when content is well-structured — with clear metadata, well-defined roles, machine-readable semantics, and a consistent authoring environment — it becomes a powerful source of context and clarity. AI tools can use this structure to reason more accurately, surface more relevant results, and even provide meaningful author feedback pre-submission. One example discussed at SSP was using AI to pre-check submissions for fit and completeness before they even reach peer review — not to replace editors or reviewers, but to help them focus their attention where it matters most.

A stronger workflow, not a shortcut

What impressed me at SSP was how many publishers are now seeing AI not as a shortcut, but as a way to strengthen their workflows. When used responsibly, AI can:

• Improve the quality of manuscript submissions
• Support faster decision-making
• Enhance discoverability and metadata
• Reduce manual overhead without reducing editorial standards

But none of this happens in a vacuum. Structured content plays a crucial enabling role. It allows AI systems to understand relationships between entities, track revisions reliably, and align author input with downstream publishing requirements. In short, it makes the entire pipeline — from authoring to acceptance — more intelligent, transparent, and efficient.

Connecting the dots

In a previous blog post, “From Chaos to Clarity: Why Structured Content Is the Key to Responsible AI in Scholarly Publishing”, I went deeper into this connection between structure and responsible AI. The key message remains the same: you can’t have trustworthy AI without trustworthy content. And that starts with how content is created.

At Fonto, we believe that structured content authoring doesn’t have to be difficult or disruptive. When done right, it feels as easy as writing in Word — but with all the benefits of consistency, metadata, and automation that structured XML can offer. That’s why we’re excited to keep supporting publishers as they embrace AI in increasingly practical ways.

Let’s continue the conversation

The SSP Annual Meeting confirmed that scholarly publishing is not just reacting to AI — it’s starting to shape it. As a community, we now have an opportunity (and a responsibility) to make sure that the tools we adopt serve the goals of transparency, reproducibility, and scholarly integrity.

If you’re exploring how structured content can support your AI journey — whether you’re a large society publisher or a smaller press looking to “level up” — we’d love to talk. Fonto is here to help you unlock the value of structure, without adding friction to your editorial workflows.

Let’s make AI work for publishing — the right way.