By David Flanagan
Rarely has the impact of scholarly publishing been more evident than over the past three years. The pandemic highlighted the importance of the research ecosystem to deliver new knowledge in a timely and trustworthy way. Now, researchers have another tool in their pockets to continue honing their craft: generative AI.
As generative AI advances, there are inherent risks that publishers are working to identify and mitigate. However, AI has immense potential to strengthen publishers’ ability to deliver trusted, high-quality knowledge.
Below are a few ways in which I think generative AI will change scholarly publishing.
1. Literature Reviews and Knowledge Synthesis Will Evolve
The integration of generative AI into scholarly publishing is poised to fundamentally change how research is aggregated and analyzed. Traditionally, keeping track of the latest results has been a manual, time-intensive process, where researchers sift through numerous papers using keyword searches to extract and synthesize relevant information.
Generative AI transforms this process by acting as an intelligent research assistant, capable of parsing vast quantities of data to produce highly personalized, comprehensive reviews. As tools incorporate web search, this also ensures that emerging findings are seamlessly incorporated, keeping literature reviews up-to-date and relevant.
The dynamic nature of generative AI-powered reviews means they can adapt to new data, offering researchers real-time insights and trends in their fields. This could particularly benefit interdisciplinary research, where connecting disparate studies is crucial but challenging.
2. New Research Avenues Will Emerge
As AI models become more sophisticated, generative AI’s potential could extend beyond information aggregation into the realm of research ideation and hypothesis generation. By analyzing existing literature, generative AI could identify unexplored areas, suggesting novel research topics or gaps in current understanding. This could lead to new research directions, particularly in interdisciplinary studies where generative AI might identify connections not immediately obvious to human researchers.
Furthermore, in the future, generative AI’s ability to propose hypotheses, especially in complex, data-intensive fields, could catalyze innovative research. This could be particularly transformative in areas like biomedical research or climate science, where intricate data patterns would require sophisticated analysis.
3. Authors Will Benefit from AI While Writing Research
Before a manuscript reaches editors or peer reviewers, generative AI can serve as a valuable tool for authors. It can be used to refine arguments, check for logical coherence, and suggest alternative perspectives or competing hypotheses. This goes beyond simple grammar or style checks, delving into the substance of the research.
Generative AI can also assist in making complex scientific concepts more accessible, aiding in effective data visualization, and providing language support, particularly for non-native English speakers. This democratization of the publication process is crucial to ensure diverse voices and perspectives in scholarly discourse.
4. New Tools to Ensure Research Integrity Will Be Deployed
The rise of generative AI presents unique challenges in maintaining research integrity. The primary concern is not with the technology itself, but with its potential misuse, including the creation of fraudulent content by papermills.
To counter this, publishers are already building better ways to verify the identities of key players in scholarly publishing (authors, editors, and reviewers, for example). At the same time, I expect generative AI to help publishers refine and enhance research integrity efforts in order to meet these challenges at scale.
5. The Role of Scholarly Publishers Will Evolve
In an environment increasingly saturated with synthetic content and misinformation, the role of scholarly publishers in curating and certifying high-quality, reliable content becomes more vital. Scholarly publishers are charged with ensuring not only that the content we disseminate is accurate and peer-reviewed, but also that it is clearly marked as such. By training and refining generative AI tools using peer-reviewed books and journals, publishers can improve the quality and reliability of the information generative AI models generate and disseminate, continuing to be that citable “gold standard” for reliable information.
Humans are always going to be at the center of knowledge creation. Generative AI has the ability to enhance the work that publishers and researchers do by automating tasks and providing unique insights to drive decision making.
Generative AI will undoubtedly have a place in the future of publishing. However, the safety and security limitations of generative AI are not yet fully understood, and legal frameworks need to catch up with, and be flexible enough to adapt to, this ever-evolving technology.
The author is senior director – generative AI product strategy, Wiley
