Trust, truth, and tech in scholarly publishing: A deep dive with Ian Mulvany

‍

About The Publisherspeak Podcast  

‍Real stories. Honest reflections. Conversations about leadership, change, and the work that shapes our world.  

‍

Hosted by Sowmya Mahadevan, Chief Orchestrator at Kriyadocs, The Publisherspeak Podcast features interviews with people making meaningful moves—inside and outside the world of scholarly publishing. From academic leaders and publishing professionals to coaches and change-makers, each episode goes beyond job titles to explore personal journeys and the mindsets behind impactful work.

‍

In this episode  

‍

Sowmya speaks with Ian Mulvany, Chief Technology Officer at BMJ Group, about the evolving dynamics of scale, identity, and trust in scholarly publishing.  

Ian and Sowmya explore why old trust systems no longer scale, how ORCID and open data can act as trust markers, and why AI might be both a risk and an equalizer. From fake papers to the promise of deeper context, this conversation gets into the complexities of building a future-ready publishing system.

‍

Dive into the conversation below or watch the episode here.

‍

Full conversation  

‍

Sowmya: It's wonderful to have you on the Publishers Speak Podcast. Thank you so much for sparing your time and being here with us today. But before I dive into this, what is that wall that you have in your office? Is that really a climbing wall that you have behind you there?

Yeah, I get asked that a lot.  

‍

Ian Mulvany: Yeah. This is my home office and last year I built a small climbing wall because it's my passion. I've been climbing for over 30 years now, and it's the only room in the house that I was allowed to do this in. So, it does dominate the background a little bit, but it creates a nice talking point.

‍

Sowmya: Absolutely.

‍

I think we are all going to talk about how we are going to climb into the future, so let's get on with it. So, Ian, you are in technology, you are in publishing. I want to hear a little bit about your story. How did you get here? You are the Chief Technology Officer at BMJ today, but what is your journey?

‍

Ian Mulvany: That's such a question. So, I started with an ambition of being an academic. And I grew up in Ireland in the 1970s, born in the 1970s, went to university in the early 90s. And back then in Ireland there was no real career opportunities for folk. You got a degree and then you left.

‍

That was kind of the route. And so, I started being really interested in physics, studying physics, going on to study astrophysics in Edinburgh, going on from there to start a PhD program in New York at Columbia University. And the academic track turned out not to be for me in the end. And after a little bit of reflection, I thought, okay, is there a role that I could do that contributes to this thing, this science that I was so passionate about that I spent nine years studying, but where I didn't have to be a scientist. And as a result, I started looking for roles in kind of support functions and accidentally, I guess ended up working in the production department for Springer in Heidelberg, working on copy editing physics manuscripts.

‍

Because I had this background in physics, I was a native English speaker and I knew the physics typesetting programming language of the day, Latex, and that got me my first role, role in publishing.

‍

Sowmya: Awesome. You know, it's so funny because the many of the guests that I've spoken to on the podcast here, they all talk about how accidentally they jumped into publishing, but how much they love it. And it's, you know, you're probably one other story like that who accidentally jumped into publishing, but you're, you know, you just love the experience and what you're able to bring to the table. So, tell us a little bit.

‍

You come from this technological background, and you've had a lot of tech experiences along the way, even though it's been in publishing. How much of this background has helped you kind of like shape where you want BMJ to head or where you want publishing to head. What is it that you bring as a philosophy into the role that you play?

‍

Ian Mulvany: I like that question. So, I'll just go back to that first role.  

The first role was working in production services, doing this copy editing, a very kind of functional role. But with my somewhat academic background, I think after about two years I kind of outgrew that role. And I was really bored, I guess, by what I was doing. And an editorial role opened up within the company, working in publishing department, being an acquisitions editor, and I moved into role in the Netherlands again for Springer, managing a couple of book series, a couple of journals, and I'm not a Salesperson. And being an acquisitions editor is kind of like a reverse sales kind of a role where you're trying to convince academics to write books for you.  

‍

And it's not my natural kind of person. So, I was not that great at that role. But it was also the era of web 2.0, if anybody here is old enough to remember that, and a time when a lot of services were coming up where you could allow people themselves to create content on these services.

‍

And it was also the time of Napster and the kind of disruption that was happening to content industries. And it was clear that publishing was a big content industry. Would it be ripe for that kind of disruption? That was the question. And what I noticed in my editorial role was we had this great connection between the authors.

‍

We had access to the full content, and we could connect the audience. So, this triangle that the publisher sat in the middle of. And I was just convinced that there must be an opportunity to create services that helped the ecosystem that weren't just built on the idea of being a content industry where all of the revenue is based on a subscription kind of model. And I was also a big proponent back then of open access, which was just starting to emerge.

‍

And with that in mind, I was looking for how I could either shape Springer to think that way or for opportunities that would allow me to work with that as an idea. But I remember just the sort of clarity of vision that given this really interesting role in the ecosystem and in the network, publishers had an opportunity to do something that few other actors in the ecosystem could do. And that was the original kind of itch. And just at that time, with that kind of motivation, combined with my realization that maybe I wasn't the best fit for this kind of job. And my long suffering boss at the time, Lisbeth Moss, must have been very patient with me because I wasn't bringing in as many books as I should have.

‍

A role opened up at Nature, working in at the time, what they called the web publishing division, where they were doing all of these little web 2.0 experiments. And it was with someone called Timo Hanay and I was based in the Netherlands at the time and I just applied on spec. I was like, well, maybe I'll apply. And then when I don't get the role it was for a product manager role, I'll ask them, well, where's the gap in my cv? And my strategy was I was going to go away and fill in the skills that I needed to work to meet that vision that I had.

‍

But that Strategy didn't quite work because I got the job, and I had to move to London and suddenly become a product manager. And that was that transition from these more editorial back-end roles into the more technology like role. And that was in 2007. And that started me on a route of being bringing some of those technical skills that I had learned through the academic period of my career into beginning to see how the actual technologies were at play in publish ecosystem.

‍

Sowmya: Yeah, no, amazing.  

And you know, I know we know each other from Elife days, and I know you've moved on from there to Sage and then to BMJ. Was it like around. I think it was around Pandemic that you moved to BMJ. Right. And Pandemic itself was like very pretty huge.

‍

I mean for the entire world, for instance, was the very first time that we had to do this. And especially for publishing. Was there certain tech decisions that you made at that time in looking at how we will shape publishing and especially for BMJ continuity and evolution?

‍

Ian Mulvany: During that time, the team that I moved in to work with at BMJ, they had already been doing a lot of work to build up the core infrastructure that allowed BMJ to survive through that period of time. So, all of the technology stack had already moved into the cloud and the tech group there had already equipped everybody in the company with VPN access into the critical systems that were needed.

‍

So by sort of a fortuitous moment of decisions that had happened in the year or two beforehand, BMJ was already well equipped to be able to work in a fully remote way when that happened. But it was kind of weird because I did accept the role before the Pandemic hit. And then my first several months were spent working having no direct contact with any of my colleagues at the time. And I don't think there was anything more than that at the time. That would have been a deep consideration around what kind of technologies were needed for the future of an organization in light of the Pandemic.

‍

But BMJ played a really critical role in supporting medrxiv at that time. And it was, I guess, the beginning of seeing what operating at a next level of scale might look like for an organization. And I think now that question of scale is one that is really at the heart of what are the challenges and opportunities for publishers as we move forward into the future. Because the scale of volume of content that's published on an annual basis means we're operating in an environment that I think doesn't have the natural and Historical mechanisms of how to think about that scale. Publishing has traditionally been a really kind of small world network, kind of an activity where you have a journal, it has senators, they know their community, it works in that kind of way.

‍

But now publishing is global. Papers are coming from everywhere in the world. And that sense that we might just already know who those communities are. Intrinsically, without the backing of a fairly large amount of technical and digital infrastructure, we can't actually operate in that way. Those assumptions that we had over the last couple of hundred years don't hold true anymore because of the scale that publishing operates at today.

‍

Sowmya: Absolutely. And I think the rate at which things are changing now is so much more, you know, it's much more accelerated. Right. There are, I mean, I don't think we plan for what's happening, but I think you can just come up with that mindset to deal with change. So, talking about change, I think AI is obviously one of the hot topics.  

We all talk about it, we all want to be prepared for what happens when AI becomes, I mean it's already here. We are all using AI in one way or the other, AI in practice. So, let's first talk about how you are probably introducing AI policies within BMJ to increase the productivity of what we are doing and what are your future? I would look at opportunities that you see for AI use in the publishing workflows or in the publishing process itself.

‍

‍Ian Mulvany: Yeah.  

So, the real steep change obviously that's happened over the last two or three years is this new class of AI, generative AI language models. And I'm going to answer your question in like two parts. I'll start by just saying why I think these models are particularly pertinent today to the industry we live in. And I'll tell you how I think about the nature of, how to think about the potential impacts that might have on our industry. So in the first part, when you train these models at very large scale, even though some people say that they are just like very enhanced versions of Autocorrect, but got, you know, but on, on steroids, being able to map all of human language and beginning to create a kind of structural representation of that language in a, in a higher dimensional space.

‍

And this is, people talk about these vector representations of language. I think what that means is that vector space has some kind of a structure. You can, you can almost project your mind up into this multi-dimensional space and if close your eyes slightly and peer into it, you might imagine that you could see valleys and troughs of structure in that Dimensional space. And that structure is like a mapping of a representation of the world itself. That's because language, the language we've created ourselves to communicate with each other, is a way to help us navigate through the world.

‍

And so, our language itself has characteristics that reflect back into our understanding of the world. Now it's kind of a shallow understanding because the way we live and inhabit in the world, it's much, much richer, but it's still there. It's not a completely abstract representation. It's good. And so, encoding that means you're encoding facts about the, about being in the world.

‍

That means these systems can actually operate with some level of competency on certain types of tasks. And so, you've got a representation in this vector space around text that can actually do things, that can do things in a way computers could never do before. When we look at scholarly publishing, every element of that value chain is deeply connected to text, from the reading of papers, to the remapping of the ideas in those papers into new shapes, through to the writing of the analysis code. And you can even begin to fold in images into this mix. So, I've seen examples in single cell biology where by just mapping and training on tens of millions of images of cells, you can begin to predict what kind of cell might emerge next in an evolutionary path of an organism, all the way through to the writing.

‍

So, there's no point in the value chain where these technologies might not have an impact. And that's why I think they're actually going to be really industry we work in. And then, and so I'm really excited by that. But I think there's a lot of uncertainty around how that's going to play out. And I think that when I then think about how we should think about the impacts and effects of this kind of technology on the future of our industry and the companies kind of companies we have.

‍

I like to think about that in, in three different domains. The stuff that we know with high certainty that's happening already and how much of an impact that might have on the work we do in the companies we work in, then there's stuff in the middle which is less certain, but where I think there is likely to be an impact in the change that might have a slightly higher overall impact. And then there's the stuff that I would classify as being less certain, but potentially very impactful or very disruptive. Now, in that first bucket, we already know that folks are using these tools to create papers, to create fake papers or to augment writing. But that's just a kind of multiplicative effect on a trend that we'd already been seeing in our industry.

‍

And so, you can make more fake papers, but the fundamental nature of that is problem doesn't change significantly. It's just, a tool in that toolbox. And the way we deal with that is to have better metadata, better understanding who the authors are, continued high quality of peer review. It doesn't change the nature of that problem, even if it's a more powerful tool in the toolbox of those actors. And in fact, I think on that side it offers more opportunity because it suddenly significantly opens the playing field to authors who have English as a second language because not only can they have a one off experience of having a much better interactive experience of translating their papers, they can spend time talking to their paper in a different language or talking to the maybe labyrinthine requirements of an academic journal and having that translated back to them in their own language.

‍

So, I think the benefits there outweigh the possible downsides.  

Sowmya: Absolutely. You know, in the publisher speak UK 2025 that we just had in me and you presented the keynote, so thank you so much for that. But I know I found it quite fascinating, and might I say that you had the highest approval rating of all of the keynotes that we've had so far, but you spoke about some very important topics that you spoke about identity and trust networks, what, you know, orchid trust markers and things like that. Right.

‍

So, and I think you gave some very specific advice to publishers and I wanted to hear it from you where you spoke about publishers have to strongly support data and code sharing as part of their scholarly record and take identity much more seriously, which continues to be a challenge and continues to be like, who will really lead that in the ecosystem and integrating with Orchid trust markers. But I wanted to hear from you, why do you think this is important and talk a little bit to us about that.

‍

‍Ian Mulvany: I think the truth is there are so many papers published, and the papers themselves are so unique and so in, in a very specific field that it's not like a ton of people can read that paper and know immediately by reading the paper whether or not the paper is truly revolutionary or indeed true. Right. It's very, you're talking to tiny communities of folk and, and so that's actually one of the reasons why fake papers get through the system because the bar to actually having a lot of people reading those papers and giving that judgment call is really hard to implement.

‍

And so, we have these proxies and we have these proxies like, oh, well, we'll have two external reviewers read the paper and we'll just assume we know who they are. And I, as the managing editor, I may have time to read the paper. Mostly I'm just going to depend on those third parties to give me the signal. And these signals are all just kind of proxies because of the complexity of that paper that we're looking at. And in a world where you know everybody, that's okay.

‍

But in a world where we now can't be certain whom anybody is, that doesn't work anymore. And that's the world we're in today. And where scientific publishing came out of the 17th century, everybody knew everybody. And we just continued with the assumption that that's how things work today. And so, our technical systems and identity systems are not at the same level as they are for almost any other web service that you might have in the world.

‍

If the web service requires transaction of some important thing, there is going to be an identity layer there which has some level of trust. And usually, we think about in the context of financial markets, I'd say the transaction of truth is at least as important for the long-term health of our societies. And yet we don't have those mechanisms of truth marker. But there are things that you can do to improve that. So, if there's a paper and the paper is making claims, it is much harder for the infrastructure of that paper to be a fake paper.

‍

If those claims, the underlying data also has to be made publicly available, if the code has to be made publicly available, if the reagents that are listed have to come from a place where those reagents have a public trail around how they're created. The more transparency we bring to the artifacts, the easier it is to verify the veracity of the claims and create more of those signals that can help us get to a place where the knowledge that we're creating is more verifiably true. Also then being able to do things like network analysis, from who the authors are, what their connections are to organizations that will stand up a claim to say, yes, that is, that is a net from this particular research organization. And I, as the research organization with the trusted marker, will support that claim. These things are eminently achievable today.

‍

And there is no downside for us trying to pursue implementing those, within our industries. And it requires, I think, that shift in mindset from one of oh, our systems today are fine because they've always been fine to the realization that they're no longer fit for purpose because of the scale that we're operating at. As I recall, I think the recommendations I said, the things you can do today are ensure that your papers, any data or code that is published or referenced in the papers are published in an open repository, get involved with the ORCID trust markers, and take identity management and identity access much more seriously. I think those are my three recommendations, as I recall.

‍

Sowmya: Yeah, I know that makes a lot of sense.  

But let me ask you a very practical question. I mean, do you see a challenge in publishers adopting of the three things that you put out there? Is there a challenge at all or is it just a matter of time? Where do you see this translating from? Okay, here are some three nice things to do too.

‍

Okay, these are like really must have. And everybody should actually go about it.

‍

Ian Mulvany: So, data deposition and code deposition is policy challenge. But there's no problems whatsoever about understanding how to implement that from an infrastructure point of view, because we've been pursuing that as an agenda for 15 years. The difficulty there is an editorial one where journals themselves will have to make the decision on whether they want to enforce that, in a hard way stance, and working with communities of, of practice, particular research communities, to make it the accepted norm.

‍

So, the challenge there is a social one, not a technical one, but we should all be pushing to make that happen. There's very little downside.

‍

‍Sowmya: I remember way back, probably about seven, eight years back, eLife mandated authors to log in to our system through Orchid. I mean, that was like, like revolutionary back then because not everybody wants to make it harder for authors to do stuff, but here you know that this was a requirement back then.

‍

Ian Mulvany: Yeah, I mean, what's the balance between what I must make you do, which is inconvenient for you, but which is better for the overall community versus things I make you do?  

Because my systems might not be as good as they should be, which is then just an annoyance. And we have a lot of that in our, in our industry as well.

‍

Sowmya: Very true.

‍

‍Ian Mulvany: On the second one on Orchid trust markers, I think that those are reasonably in a good enough shape that we should all be investing and investigating how we implement those in our workflows. But I think that there might not be as much infrastructural support for, for them in publishing systems today because they're relatively new to our ecosystem. And so, I think the adoption barrier there is slightly higher. And then on the last one, how we change identity systems, that is a totally open question. Because there are many technical implementations that any given publisher could look at things like know your customer protocols.

‍

But it comes with a change in an understanding of what are the right kinds of places to put that trust question. For example, if you're an author who has published 15 times with a publisher and you're on an editorial board, why does that publisher not know this and offer you a faster route to publication vs you're an author who hasn't been seen before by that publisher or by that funder there? You might say, well, we might put higher trust barriers in place, but even those two things require an ecosystem or infrastructure that itself can be flexible and can understand the history and metadata of all of the interactions have happened between that identity and the system. That's non trivial to do. And then secondly, what are the policies that you put in place where you say, well, if it's exactly this number of interactions, you go from trust heavy to trust light.

‍

And who makes that decision? And then that's a social question that we haven't had much time talking or considering. And so, I think, the challenge with implementing anything new in our industry is there is a natural resistance and conservatism. And often what happens is someone does it first and everyone else goes, oh, that's quite interesting, I'll follow what you've done. But getting to be the first person to do it can be a real challenge in the industry.

‍

Sowmya: True, but I think, yeah, so coming to that on that note, right, I do want to ask you about what do you think is the future of publishing? I know we had some conversations about this at the Publishers Speak conference. There were actually questions around about relevance of a publisher, in the Tomorrowland. Right. So would people just have their signs on medium and, and here their LinkedIn et cetera, et cetera.

‍

What is the relevance of publishing? But then I'm also hearing about, you know, hey, academia, which kind of wants people to publish and be, you know, published. Scholar is changing at a much slower pace than publishing. So we're still going to require these publish. What are your thoughts?

‍

These are the things that I'm hearing from, from different voices. But what are your thoughts about future of publishing? Future of this industry? What are you worried about? Also, what are the opportunities that you see and what keeps you quite excited.

‍

Ian Mulvany: So, no small questions.

‍

‍Sowmya: No small questions. World peace. We have to solve world peace.

‍

‍Ian Mulvany: Yeah, yeah.

‍

Have been known to make predictions in the past and be wildly wrong. So do you remember, a thing called Google Wave that came out Many years ago, I don't know if you remember Google Wave, it was this, no, it was this product that came out from Google and it was at the time when Google would launch a new product, but it was top secret, but you'd get an invite finally and it was back just when Gmail was beginning to be released a bit more. And Google Wave was this real time collaborative platform. You could both be in a document at the same time. It was before code editing existed in Google Docs and that's actually the inheritor of that functionality.

‍

And you could have bots and agents that would sit in the document and wait for something to happen. And I remember being, this is the future. We'll have documents where agents sit and co-write with us and then that will be the all publishers need to change them. And I was giving talks about this and creating died. It totally died, right?

‍

It's like I was totally wrong about that. So, I can give you prognostications about the future, but I'm pretty bad at being accurate about them, but that won't stop me. So like going back to the thing about AI, right, Like I think the area which is potentially most radical to think about is could you have bots or agents that are competent enough to write and analyze and do scientific work, in which case do you have a future in which the author themselves becomes less like they are today in how they are contributing to what that artifact is? But humans still are really curious and obsessive and really like getting credit for things. And so that driver of how beautiful the world is and how rich the experience you have of learning about the complexity of the world and uncovering something new about the world, I don't see those things going away as drivers of human curiosity.

‍

Even if we have another type of intelligence in the world that doesn't mean remove or replace the drivers that we ourselves have as humans. And so I think there'll always be that curiosity and driver and there will always be problems that will need to be matched by that. And so, one of those things that I think won't change over time will be the drive that people will have to understand and a system that will help them be rewarded for when they find new insight into the world. And today the publishing ecosystem, for all of its ills, acts as the game in which those claims are made, assessed and ultimately awards are granted on the basis of that, be they just career progression or career stability or prizes or what have you. So, I don't think that fundamental thing changes.

‍

And so if that's the case, then I think as publishers, what we should be doing is doubling down on what creates high integrity for the scholarly record and maybe opening up to the door to saying, well, now we have a different kind of an author or a different kind of a contribution, which may be semi-autonomous or fully autonomous, but it has to play in relation to the. The wants and needs and motivations of humans in that system as well. And I think as well that as technologies become more powerful in being able to see context or see meaning or see trends in how we might label corpuses of text or how we might categorize them, we have an opportunity to build systems that provide higher context to the people who are operating within them. And so right now, if someone is writing a paper and that paper has certain claims in the paper, what might it be like when you interact with the publishing system, that the publishing system comes back and says, okay, you've referenced these works, but here's everywhere where a similar claim happened in a field that is far from the field that you currently work in. Would you like to know about that?

‍

Just help you deepen your thinking as you go into to write this. Or if there is a statistical method used in a paper, wouldn't it be nice if the publishing process was able to automatically run that on the data and then run it on every example of data like that that had already been published, just to help give the author a sense of how significant their result is in a meta-analysis that could already happen for them at the moment of publication. And those are examples of imagining how our technologies could be applied with the window of creating richer context for the humans who are participants in it, with a view to lowering the burden of complexity to increase the value of thought that happens in the system. I think that's a really interesting opportunity space to play in, though. That's like probably many years down the road from where we are now.

‍

But equally, all we have in front of us is time, so why not think in that way? Right?

‍

‍Sowmya: Yeah. No, I love what you actually said. Lowering the complexity to increase the context.

‍

I think that totally kind of describes what AI or technology can do for you. It's getting you to be more in tune with the core of what is being discussed rather than worrying about all the fluff around it. Right? So, yeah, I really enjoyed this conversation today, Ian. I thank you so much for all your insights.

‍

A lot of takeaways from me here, and I'm sure our audience would have really enjoyed all your, you know, gazing into the crystal ball as well. As all the insights that you had to share. But yeah, I really enjoyed our conversation. Thank you so much for being on the Publishers Big podcast.

‍

Ian Mulvany: It's been a real pleasure.  

Thank you for the questions. Yeah, let's work on making this happen.

‍

‍Sowmya: Absolutely. Thanks a lot, Ian.

‍

Explore The Publisherspeak Podcast on Spotify here.