AI and academic publishing: What does the future hold for authors, readers and publishers?

AI and academic publishing: What does the future hold for authors, readers and publishers?


In this webinar held on 13th May 2024, we analysed what the AI revolution means for you, as authors, editors, reviewers, readers and publishers, and considered possible responses from policymakers, following a recent report and policy recommendations to the European Commission. 

Key messages

The webinar was organised by the Academia Europaea Cardiff Knowledge Hub as part of the Scientific Advice Mechanism to the European Commission.

The Chair, Ole Petersen, welcomed the panel of speakers, and provided the context for the webinar.

The opening speaker, Professor Nicole Grobert, outlined the main findings and recommendations of the Scientific Opinion, Successful and timely uptake of AI in science in the EU. The European Commission had asked for advice on how AI could be taken up in science in a responsible way, whilst maintaining the EU’s global competitiveness. In response, the Scientific Opinion describes how an appropriate environment and infrastructure can be created that facilitates increased uptake. Key recommendations include the provision of access to high-quality data, tools and state-of-the-art AI systems; appropriate funding mechanisms; the creation of a distributed European AI institute and an AI in Science Council.

Dr Anita de Waard, representing the publisher Elsevier, presented an overview of the use of generative AI in academic publishing. One of the principal concerns of publishers is the lack of training of AI bots on high-quality scientific outputs, leading to risks of bias in content. Research integrity is already an area of concern, made worse by the proliferation of fake papers. Moreover, the environmental impact of the huge computational power required for AI is concerning. At the same time, AI brings significant opportunities. AI has the potential to increase the quality and creativity of research and scholarship. The tasks of writing, editing and reviewing can all be improved by the use of AI. AI can enable new forms of scholarly communication, as well as boosting the communication and impact of research. AI can also help detect bias, improving equity and inclusion. Elsevier has developed a set of responsible AI principles. It is working with the AI community to identify fake papers, and to detect whether a paper has been written by a human or by AI, as well as conducting plagiarism checks. The abstract/indexing database, Scopus, now incorporates AI tools into its frontend. Elsevier is also looking into ways of using AI to produce policy briefs.

The next speaker, Professor Alberto Melloni, addressed the use of AI in the humanities. He gave examples, such as the use of AI to map critical editions of major works held in different libraries, and the use of AI to analyse fragments of ancient writings. AI provides humanities scholars with opportunities to bring technology to the fore in their research.

The final speaker, Professor Paul Groth, highlighted some of the key findings that emerged from the Evidence Review Report produced by SAPEA, which informed the Scientific Opinion. AI can be used across the research life cycle, from the generation of hypotheses from the literature, to the automation of workflows, to the production of literature summaries. In short, AI can be an accelerant for many research-related processes. However, there are challenges.  Many AI tools are held privately, and there is a need to combat the perception of low-quality information.

Dr Kiera McNeice, represented Cambridge University Press, supported many of the statements made by other members of the panel, emphasising the importance of validating the integrity of scientific research. 

The discussion that followed the panel presentations ranged across a number of areas. These included scrutinising and checking submitted manuscripts; comparisons in task performance between AI and humans; tackling bias within datasets; inequalities in access to AI; the origins and consequences of fake papers; the relationship between Open Access and AI; and questions of reproducibility.

Further details are in the recording and slides.


Watch the webinar


Slides




Updated 5th May 2024. For further information please contact AECardiffHub@cardiff.ac.uk



Share this page: