Publishing Evolves in a Connected World
By Ann Griswold, NYAS Contributor
In the days before artificial intelligence mined obscure gems from the scientific literature; before preprint servers posted study results without pausing for peer review; when social networking meant cocktail conversations at the industry conferences — science publishing looked very different than it does today.
“When I started, authors submitted paper manuscripts produced on a typewriter,” says Douglas Braaten, PhD, chief scientific officer responsible for the Academy’s science journals and books, and editor-in-chief of Annals of the NY Academy of Sciences. “Happily, not everything stays the same.” But as technology continues to reshape academic publishing he says, “it’s useful to think about the things that do.”
In many ways, the process from paper submission to publication has not changed much in 40 years. Researchers still submit papers to journals. Papers are peer reviewed. Journal editors share the reviews with authors. Revisions take place. And then submissions are transformed into carefully copy edited, typeset manuscripts. This system has survived because it works, Braaten says.
But critics have long called for change. Some note that the peer review process can stretch on for months or even years. Others point out that the $25 billion academic publishing industry is dominated by a handful of major players who make a profit from the public investment in research.
Braaten and others don’t deny that the system could be improved, but they say the situation is more nuanced than critics suggest. “I’d love for the industry to be less concerned with profit-making,” he says. “But there are some fundamentally important and useful things about peer review and having vetted, polished papers published in journals with global footprints.”
Still, technologies aimed at tweaking the process have increasingly flooded the market — and are gradually transforming the business. “It’s hard to keep track of these innovations because there are so many of them out there,” says Steven Ottogalli, Publisher, Life & Physical Sciences at Wiley. “Start-ups are coming online and impacting every part of the publication process, affecting every aspect of the value chain that used to lie solely with the publishers or with the academic societies.”
Preprint servers, for example — long a standard in the math and physics communities — are gaining in popularity with biologists. Scholarly collaboration networks are connecting researchers from diverse fields and distant locations, allowing them to share their findings in real time. Artificial intelligence-based search tools are tipping off scientists to papers they might otherwise overlook — creating new synergies. Novel technologies such as blockchain aim to increase accountability and transparency in the review process by encoding each article with a record of its origins, revisions and peer reviews. And younger generations of investigators are replacing static figures with embedded multimedia and interactive data.
Collectively, these innovations can bring the world to a laboratory’s doorstep and allow siloed projects to spread in new, unexpected directions. But how will these new technologies fundamentally change the traditional model of scientific publishing?
“There’s such a sociology surrounding scientific publishing,” Braaten says. “Think of what it means for grant funding, and tenure evaluations. And for what it means for the careers of young investigators when they publish in a top-tier journal. One would have a very hard time replacing all of these significant benefits with changes just in technology.”
Yet it’s hard to deny that the field is in transition. “I’ve been in this business for almost 20 years, and things have changed so drastically,” says Ottogalli. “Who knows what it will look like in another decade.”
Variations on a Theme
Most of the innovations Ottogalli mentions are variations on the theme of open access, a business model that shifts the financial outlay from academic institutions to authors and funders (such as the Wellcome Trust) by replacing subscription fees and paywalls with open access license fees and free access to published papers.
Critics complain that subscription-based journals restrict access to publicly funded research by creating subscription paywalls. This, they say, forces the very academic institutions that produce the findings to pay for access to the published work. Open access could potentially fix that by providing immediate public access to papers upon publication.
“But this ‘fix’ doesn’t address the whole story,” Braaten says. “If by open access, you mean access to the information, there are a lot of ways now that one can access all published research, publicly funded or not.”
For example, most journals allow authors to post the submitted version of their manuscript on a lab website or, after peer review and acceptance, on the post-publication websites such as PubMed Central. And with the advent of preprint servers, Braaten says, there’s a huge amount of accessible information for free. “If someone wants access to a paper, it’s often available on one of several sites — just in a different format than one finds in a published journal.”
Subscription paywalls don’t keep science from the people, he says. Rather, they provide publishers and journal owners the funds required to produce published peer reviewed and polished papers on websites in HTML and PDF form, and in print journals. “The typeset published version — not the actual science in the article — is the thing that’s owned by a journal or publisher — it’s the product of their work. I think people may not be aware of this distinction,” he says.
Access to Published Science in a Connected World
One upside to all publishing models — including open access and pre- and post-publication servers — is that published papers are available for such things as AI text mining. In turn, this can improve discoverability within disparate disciplines, for example, ecology and economics. “AI will help humans make connections where they didn’t think of making connections before,” Ottogalli says.
The AI tool IBM Watson, for example, can search published content and find papers on climate change that a researcher might be interested in — and then make connections to other papers that might unexpectedly support work on climate change.
In the meantime, researchers are finding ways to share their findings outside of the traditional publishing process.
“Scholarly collaboration networks, which are like Facebook for researchers, are providing greater opportunities for working together,” Ottogalli says. ResearchGate and other scholarly collaboration networks (SCNs) build ties among researchers in similar and disparate fields, and can put relatively obscure labs in developing countries in touch with larger, well-funded ones abroad.
“The international collaboration piece is very important,” Ottogalli says. The publishing landscape is dominated by Western Europe, the U.S. and China, so “sites like this open up doors for researchers who may not be known to researchers in Europe or the U.S. At the end of the day, I think SCNs are a valuable tool in advancing science.”
But There’s a Downside, Too
ResearchGate and other SCNs can be aggressive in encouraging authors to upload versions of their published PDFs in violation of copyright. "This is clearly wrong," Braaten says. Publishers have routinely issued take-down notices, informing the authors and ResearchGate that they must remove the content because it is in violation of publishers’ policies.
Echoing Braaten, there’s another way to provide access to research findings says Ottogalli: “preprint servers.” In a 2016 policy forum, Science lauded the advantages of preprint servers for authors, journals and funders. These sites expedite publication and offer a forum for sharing new tools or negative results, potentially accelerating the pace of research. It’s also possible that preprint servers could help weed out questionable scientific papers in the pre-peer review phase, when other researchers comment publicly on the study. “Some authors may value the feedback before the paper is submitted to a peer-reviewed journal,” Ottogalli says.
The concept is rising in popularity. Some journals have launched their own preprint servers for papers under review. And a few major federally funded programs require their investigators to post preliminary findings to the servers.
“We’ll see really interesting advancements in the next few years,” says Ottogalli. “It’s a time of big change.”