“We can help make science faster, more accessible and more trustworthy”

Springer Nature stressed its leading position in the development of academic publishing infrastructure by investing, in the last three years, nearly half a billion in tech.

Joyce Lorigan, Group Head of Corporate Affairs at Springer Nature, explains what’s in the box, why Springer Nature uses AI extensively, and how all this is going to pay off – for science, for the publisher and for the broader world.

How come publishers such as Springer Nature started to embrace the once 'public enemy No. 1', that is artificial intelligence?

I do not think we're talking about an enemy.

Imagine a world without groundbreaking technology like smartphones. Or the internet, which was invented by scientists for scientists. These innovations exist because scientists asked questions and pursued answers. Researchers have historically been at the forefront of innovation, quickly adopting technology to aid their work.

As part of the scientific ecosystem, this also applies to us. Technology is fundamental to our mission of serving our communities. Over our 180-year history, it has always been a crucial element, helping us to better support researchers, peer reviewers and editors. Technology has allowed us to transition from print to digital earlier than many other content industries and is pivotal in our shift to making research immediately available to all via open access.

We have been successfully leveraging AI for over a decade to transform the way we support researchers. We believe that this technology holds great potential to further enhance discovery, provided it is managed carefully with human oversight. This is essential as we confront global challenges such as climate change, where the world looks to scientists for solutions. For science to meet these expectations, it must become faster, more accessible, and more trustworthy. AI could be a game-changer in achieving these goals.

Plagiarism and fraud are rife in scholarly publishing. Wouldn't it be time for a new ethos instead of tools which are known for their ability to support researchers' attempts to cheat?

I do not think this is a matter of choosing one over the other. We need both a commitment to ethics in science, including with regard to the use of AI, and robust mechanisms to detect scientific fraud. Paper mills and other fraudulent practices pose an increasing threat to the trustworthiness of science. Protecting truth and integrity is becoming ever more critical for both scientific progress and the practical application of research.

The scientific record is rapidly expanding, in part driven by the need to address the global challenges we are facing. Given the pace of this expansion in relation to the number of people available to review the science, relying solely on the human eye is not enough to validate research and prevent problematic content from entering the system. In 2023, Springer Nature handled more than 1.8 million research submissions, which equates to almost 5,000 every single day. We published over 420,000 articles, slightly less than a quarter of those received.

"By investing in technology and research integrity, we are building trust in science."

With the aid of AI-driven tools, we can enhance our ability to detect plagiarism, the misuse of large language models for creating fake science, image tampering, and potential conflicts of interest among authors and peer reviewers. By investing in technology and research integrity, we are building trust in science.

Collaboration among the members of the scientific community is key because we are all in this together. When integrity issues arise in research publishing, it can affect trust across the whole industry. This is why we are working with other publishers to foster confidence for the future, through organizations such as the STM International association, partnering with its Integrity Hub initiative.

What is the general idea and strategy behind Springer Nature's AI-related initiatives?

AI has the potential to make processes simpler, faster and better. We believe that it can significantly accelerate the pace and enhance the quality of research, allowing us to serve our communities more effectively. But there is a significant gap between usage and trust: A recent study published by Oxford University Press indicates that three-quarters of researchers already use AI routinely, yet less than 10% trust big technology companies for privacy and security. A Nature survey comes to a similar conclusion: scientists are excited and concerned by the increasing use of AI tools in research at the same time.

This is where we come in: As a scientific publisher, we hold a unique position at the intersection of discovery and application. By collaborating with tech entrepreneurs to develop and utilize technology that supports research and the publishing process, we can help make science faster, more accessible and more trustworthy.

"Relying solely on the human eye is not enough to validate research and prevent problematic content from entering the system"

We put humans and our communities at the center of our decision-making: when using AI, there is always a human in the loop who takes responsibility. Our AI strategy focuses on expanding our technical expertise while remaining committed to using technology in an ethically responsible way.

You claim to support not only scholarly writing but the preceding research, too. Why do you think this is publishers' business?

Our primary purpose extends beyond simply publishing scientific journals and books. It is to promote research findings that help address the pressing challenges of our time by advancing discovery. We want to enable researchers to focus on their research by streamlining the writing and publishing processes.

So why not extend our support to help them develop and disseminate their findings more effectively across the research cycle? For example:

Providing Platforms: Why not offer a platform like Protocols.io as a secure space for developing and sharing reproducible methods, thereby avoiding redundant research and paving the way forward for an open science future?

Language Support: Why not assist non-native English-speaking scientists with automatic translations to broaden their impact?

Manuscript Assistance: Why not help academic book authors save time when writing, making their knowledge available more quickly while maintaining high quality?

As a publisher, we work closely with researchers and understand their needs. With AI, we can equip them with a variety of tools and platforms specifically designed to optimize their daily workflows and enable them to become more agile and efficient.

 

Give us a rough idea of how much Springer Nature invested in AI tools for the support of scholarly publishing.

Artificial intelligence constitutes a significant portion of our investment in technological expertise. In 2023, we increased our technology-related spend, including personnel costs, to more than €173 million, taking our total investment in these areas to over €470 million since 2021.

"We can help make science faster, more accessible and more trustworthy."

This has helped us reach significant milestones for our researcher-centric tools, solutions and platforms. We’re using AI and machine learning to power Snapp, our new article processing platform, which exceeded 1 million submissions last year and which is a good example of how we are making publishing simpler and more rewarding. It has also helped us improve our ability to protect Research Integrity by developing advanced AI tools like Geppetto, which scans submissions for suspicious text, meaning we can now turn away hundreds of fake science submissions every month. With Protocols.io and Slimmer AI's Science division, we have also successfully integrated recently acquired AI and technology-related startups into our company.

 

LLMs must be trained on existing media content. Will Springer Nature's AI be trained on Springer Nature's content?

At the moment, our focus is on how we can best use AI internally to help researchers get published faster and find relevant content more quickly, make sure research can be trusted and to reduce barriers to equity. We have been using AI for over 10 years already in support of these goals. Regarding LLMs, we do not currently license our content to third-party LLM providers for the purpose of training their models.

It is, though, a fast-moving area, and we are always open to things which will improve the experience for our customers and the wider research community. Therefore, should this change, we will apply the same considerations we have to date – does it benefit our customers and the wider research community? Is there a ‘human in the loop’? Does it meet our AI governance principles of fairness, transparency, accountability, and respect?

 

Are we standing at the threshold of a new era of scholarly publishing with a high output and a high quality at the same time?

Between 2011 and 2021, global R&D investments grew from $1.4 trillion to $2.2 trillion, with the number of published scientific articles increasing from 1.6 million to 2.6 million by 2023. These are positive developments, and yet the pressing challenges faced by our planet necessitate even swifter and more inclusive scientific progress, all while maintaining high standards of quality. It is crucial for research to be reliable so that other scientists can confidently invest their time and resources to expand upon it. We believe that technology will be the catalyst for this enhanced and necessary progress.

Joyce holds the distinguished role of Group Head of Corporate Affairs at Springer Nature, a leading global science communications company, where she not only contributes to the Executive Team but also spearheads sustainability initiatives. Beyond her corporate endeavors, she has chaired the esteemed London-based business alliance, Urban Partners, and serves as a trustee for the esteemed Marine Conservation Society. Before her tenure at Springer Nature, Joyce amassed extensive experience in senior communications and marketing positions with renowned global brands. Her professional journey uniquely weaves together corporate leadership with entrepreneurial ventures, including a foray into travel technology and the establishment of her own communications consultancy startup.