AI/ML is Transforming the Scholarly Publishing Industry

Posted on : January 30th 2023

Artificial intelligence (AI) is significantly changing how researchers interact with their audiences while also transforming the scholarly publishing sector to become more technologically focused. The surge in research output, combined with the increasing demand for peer reviewers, has prompted the adoption of AI in scholarly publishing. Despite the fact that human input is considered necessary for the authoring of publications, the contribution of AI is slowly but steadily gaining ground.

Application of AI in scholarly publishing

AI can contribute to a shift in how published articles are seen and accepted. It is essential that publishers train an artificial intelligence system on a large amount of research data and algorithms in order for it to be efficient and predictive. With several publishers already fostering AI systems, the publishing industry is poised for significant expansion in the coming years. As a result, publishers have already started to use Natural Language Processing (NLP) to deliver semantic enrichment and content recommendations to users. Additionally, they are using XML to explore new business opportunities, new audiences for their journals, and to improve the quality of their existing research publications.

The use of AI and NLP in the peer review process is becoming more common among publishers, who are using it to retrieve key terms for uniqueness and relationship mapping, matching papers to journals and reviewers, assess manuscripts, check for language quality, and even detect plagiarism. Applied at varying stages in the content life cycle, intelligent automation of workflows minimizes the time it takes to publish content, enhances editorial quality, and improves author experience.

The scholarly publishing industry has seen the emergence of a number of AI-assisted solutions. AI algorithms generate summaries of scientific articles and transform them into plain-language texts, press releases, and news reports. Accurate and sufficient information retrieval is key in evidence-based research publications. Semantic technologies can facilitate transparent and effective data extraction techniques. AI's function in scholarly publishing is intended to be multifaceted, ranging from detecting plagiarism issues to forecasting the projected citation impact of a yet-to-be-published article. Writers, editors, authors, and publishers should take advantage of AI, ML, and NLP in order to facilitate the rapid and accurate distribution of scientific information to contribute to the advancement of humanity as a whole.

The rapid development of AI over the past decade is largely attributable to the increased transparency of the field on the part of both academia and industry. However, that may soon change as researchers race to create more powerful AI. Progress in AI could stall if researchers became less willing to share their findings.

Since its release in November 2022, OpenAI's ChatGPT, an artificially intelligent chatbot, has been the subject of much interest. It uses the GPT (Generative Pre-trained Transformer) architecture and has been trained using a huge corpus of internet-sourced text. The model can produce human-like text by recognizing patterns between words and using those patterns to speculate on what to write next in response to a written prompt or set of instructions. The question of whether or not these systems engage in plagiarism of authors is open to debate.

This year, officials chairing the International Conference on Machine Learning made headlines by prohibiting researchers from submitting papers that included text generated by large language models and tools like ChatGPT. This decision has sparked a discussion about the value of AI-generated text in the academic community. The announcement sparked debate on social media, with researchers and academics in the field of artificial intelligence both supporting and criticizing the policy.

Legal issues surrounding AI-created works are a constant source of frustration for those working in the AI field. However, if we're not careful, the very thing that seems to mark the golden age of AI may in fact mark its end.

Optimizing content discoverability

AI has the potential to significantly improve processes throughout the scholarly publishing workflow. It enhances the ability to process the rapidly increasing number of manuscripts that are being submitted, as well as the accuracy of the data and findings that are being produced. In general, it accelerates the dissemination of scientific breakthroughs and advances scientific discovery.

Publishers can use AI/NLP to improve discoverability, workflow efficiency, user engagement, and output quality, as well as optimize their marketing and sales strategies. Intelligent machine learning algorithms can help auto-taggers in accurately tagging articles and highlighting incorrectly assigned tags. This simplifies the implementation and management of taxonomies for publishers. Additionally, publishers can leverage NLP to retrieve and rank key terms from various sections of their content. This would eventually improve content categorization and, consequently, content discovery. Artificial intelligence can be used to create algorithms to search for trending topics, institutions, and reviewers in their field of expertise, among other things. Similarly, machine learning can be used to separate labels from their associated images and extract sub-figures and captions from compound figures.

Enhance author experience

AI/NLP should be used primarily to enhance the author’s experience, rather than to reduce costs. Using AI, ML, and NLP, the industry should strive to improve the author experience while also streamlining and increasing the efficiency of various workflows.

The opportunities are endless, and these solutions are assisting the industry in addressing the challenges posed by shifting pricing models and economic conditions.

While larger organizations may invest in technologies that are financially viable only for their size, smaller publishers can take a digital approach to reduce production and delivery costs. Online publishing, facilitated by digital technologies, benefits publishers by removing the expenses associated with producing and distributing a printed journal. Integrating AI into online publishing platforms can help create a highly advanced and standardized publishing landscape that requires minimal human interaction.

Conclusion

While it is difficult to predict how AI will affect academic publishing in the coming years, it is clear that it will have a significant impact. Although we do not anticipate AI taking over the publishing process entirely, there are several aspects across the workflow where AI will dramatically enhance the delivery, accuracy, and comprehensiveness of the published research. Constantly evolving AI tools can enable publishers to significantly improve the processing and distribution of scientific research and discovery. However, human review and reasoning are still required in places that are beyond the capabilities of AI.

Straive is actively researching and implementing innovative ways to make upstream processes more efficient, by integrating AI-powered gatekeeping as a value add to enhance the author and editor experience.

Visit our website to know how Straive can help enhance your publishing experience.