The Society for Scholarly Publishing released its Pulse Check report in January, surveying 563 professionals across scholarly publishing about their AI usage, concerns, and expectations. It's the most comprehensive snapshot we have of where the industry actually stands — not where conference keynotes say it stands, but where the people doing the work say it stands.

The headline numbers are worth knowing. The gaps between them are worth understanding.

The adoption paradox

The survey confirms what anyone working in publishing already suspects: AI adoption is widespread and accelerating. Large majorities report using AI tools in their daily workflows — for manuscript screening, editorial correspondence, metadata generation, and increasingly for peer review logistics.

But here's the paradox. While adoption is high in practice, institutional comfort with that adoption is low. Publishers are using AI tools faster than they're writing policies to govern them. The gap between "our staff uses AI daily" and "we have clear guidelines for AI use" is significant, and it maps almost exactly onto organizational size. Large commercial publishers have formal frameworks. Mid-size society publishers are drafting them. Small publishers and university presses are largely improvising.

This isn't a criticism — it's a structural observation. Writing AI policy requires legal review, board approval, and staff training. Using ChatGPT to draft a rejection letter requires a browser tab. The friction is asymmetric, and the gap is growing.

The disclosure problem

Separately, a COPE analysis published in the Scholarly Kitchen found that 84% of publishing professionals report using AI tools in their daily work — but most authors still don't disclose AI use in their manuscripts, even when journal policies require it.

This creates a compliance gap that no one quite knows how to enforce. Do you reject manuscripts where AI use is suspected but undisclosed? Do you develop detection tools — knowing those tools have their own reliability problems? Do you accept that disclosure requirements are aspirational rather than operational?

The industry is currently doing all three, depending on who you ask.

What the numbers don't say

Surveys tell you what people report. They don't tell you what people do when no one's surveying them. The real story of AI in scholarly publishing is happening in the space between official policy and actual practice — in the editorial offices where AI tools are quietly making peer review faster, in the production pipelines where automated screening catches (or misses) paper mill submissions, and in the licensing negotiations where publishers are trying to figure out what their content is worth to an AI company.

We'll be tracking all of it. The SSP Pulse Check is a snapshot. What this site does is the time-lapse.