Experts debated on November 7, 2025 the future of academic publishing—from reforming the current system to building something entirely new. Watch the full recording and explore the key arguments below.  

→ Watch the debate here
 

A System Under Pressure

Scholarly publishing stands at a crossroads. While academic journals have long structured how research is communicated, evaluated, and rewarded, growing concerns about reproducibility, bias, incentive misalignment, and information overload have sparked calls for more radical change.

In keeping with the Oxford-style debate format, the speakers presented deliberately sharpened positions. Moderated by Jess Rohmann, the discussion nonetheless revealed broad agreement on the failures of scholarly publishing, but a fundamental divide over the solution: reforming the system from within or replacing it entirely.

Bernd Pulverer argued for reforming journals from within, while Alex Freeman made the case for replacing them entirely with a modular, open publishing infrastructure. The debate began with formal presentations, followed by audience voting and an extended, probing discussion.
 

“We Can Fix This from the Inside”: The Case for Reforming Journals

Bernd Pulverer, Head of Scientific Publishing at EMBO Press

Pulverer’s argument was grounded in pragmatism: journals are imperfect, but they remain the most scalable, structured, and independent mechanism science currently has for quality control and synthesis.

He emphasized that journals do far more than brand articles. They organize research into interpretable units, provide narrative context, and enable interdisciplinary reading—functions that many researchers, clinicians, and policymakers rely on. Crucially, he argued, journals can evolve without abandoning their core role.

“We are trying to activate postdoc-level referees to look into the source data that we publish with our figures—people who actually have the time and expertise to dig into the data itself.”

Pulverer acknowledged deep problems: selective publishing, reproducibility failures, and incentive systems that reward prestige over rigor. But rather than dismantling journals, he proposed layering granularity and transparency around the research paper—for example through source data publication, method validation, and AI-assisted quality checks.

He also stressed the value of journals as independent filters, particularly in matters of research integrity:

“It’s actually extremely useful to have a nozzle that is independent of research institutions—especially when integrity investigations are needed.”

Drawing on his experience in the life sciences, Pulverer argued that fields like biology and medicine—large, distributed, and data-messy—still require structured editorial oversight. Incremental reform, not replacement, was his prescription.

“The Format Is Rotten to the Core”: A Call to Rethink Publishing Entirely

Alex Freeman, Founder of Octopus

Freeman’s argument went deeper than critique. She challenged the very format of the journal article, describing it as a legacy structure optimized for narrative storytelling rather than honest, reproducible research.

She began by asking why researchers publish at all: to communicate work, to enable trust and assessment, and to be assessed themselves. In her view, journals fail at all three. Articles rarely contain enough detail for other researchers to fully understand what was done or to replicate it. Negative results, pilot studies, and abandoned ideas disappear. Large datasets, code, and non-textual outputs are squeezed into PDFs that cannot meaningfully represent them.

“To create a narrative, you prune away all the side issues—all the things that didn’t work. And that’s not how research actually works.”

Freeman warned that this narrative pressure nudges researchers toward questionable practices—hypothesizing after results are known, removing inconvenient data points, or p-hacking—especially when journals double as career gatekeepers.

Her alternative, implemented in Octopus, replaces articles with linked, modular publications: methods, data, analyses, interpretations, and reviews are published separately but connected. Peer review is open, transparent, and treated as a scholarly output in its own right. The platform allows researchers to publish all details so that others can fully understand and build on their work, addressing the reproducibility and transparency gaps left by journals.

Her most striking metaphor addressed incentives:
“At the moment we’re incentivized to build the tallest towers with our articles—putting flags on the top and getting rewarded for height. People aren’t really looking at the quality of the brickwork or the foundations.”

In Octopus, she argued, value shifts to the quality of each contribution—each brick of a tower—allowing trustworthy knowledge to emerge collectively over time.
 

When the Floor Opened: Where the Debate Got Concrete

The audience discussion pushed both speakers beyond principle into practice.

Who filters the flood?

With AI-generated papers and exploding volumes of data, filtering emerged as a central concern. Freeman argued that openness enables community-built tools, automated checks, and collective moderation—comparing the model to Wikipedia or GitHub. Pulverer countered that voluntary oversight alone is unreliable and that journals already shoulder this burden at scale, albeit imperfectly.

Reproducibility vs. Novelty

Both speakers agreed that reproducibility is essential—but under-rewarded. Pulverer cited evidence that only around half of bioscience studies are easily reproducible, calling this “extremely disturbing.” Freeman emphasized that the solution lies in shifting research incentives:

“What should excite us is not always going further into the unknown, but asking whether what we’ve done is as robust and solid as it can be.”

Time, incentives, and careers

Repeatedly, the discussion returned to incentives. Freeman argued that researchers appear to “lack time” for open practices only because those practices are not rewarded. Pulverer agreed that institutions and funders must embed new expectations into hiring, training, and evaluation—but cautioned that selection pressures in science are unlikely to disappear.

Different disciplines, different cultures

Physics and astrophysics were cited as examples where open, collaborative practices work well—largely due to shared data infrastructure and strong community norms. Both speakers agreed that biology’s scale, complexity, and competitive structure pose different challenges, but also that cross-disciplinary lessons are essential.

Reform or Replacement?

Despite their opposing positions, the debate revealed striking common ground. Both speakers called for greater transparency, better incentives, professional recognition of peer review, and cultural change. Where they diverged was on where to anchor reform: Pulverer argued for evolving journals as the organizing core; Freeman insisted that only a new, modular infrastructure can realign science with its own values.

While the audience remained divided on how to fix scholarly publishing, the discussion made it clear that everyone agrees the system needs change. 

For anyone concerned with how knowledge is built, trusted, and rewarded, this debate offers not just critique, but competing—and complementary—visions of what scientific publishing could become. Watch the full recording to explore the arguments in depth. 

The Event was organised by the Einstein Foundation Award and the Berlin University Alliance, as part of the Berlin Science Week 2025.
 

→ More events on research quality