Springer Nature uses generative AI to publish academic book

Springer Nature has become the first publisher to create a whole new academic book by empowering authors to use GPT as part of the integrated workflow. Developed during a ‘Hack Day’ in the Spring which brought together authors, editors and experts from across Springer Nature, the German-language book Einsatzmöglichkeiten von GPT in Finance, Compliance und Audit (Applications of GPT in finance, compliance and audit) has now been published. It took less than five months from inception to publication – about half the time normally taken.

The team wanted to see whether generative AI tools can help authors save time when writing their manuscripts and make their knowledge available more quickly, while maintaining a high level of quality. The experiment also showed that the technology can lower barriers for people who, for various reasons, find it difficult to write and would not be able to author a book without assistance. However, it became clear that authors with their expertise and high-quality editing continue to play an essential role in ensuring high quality content.

The process was as follows: 

1. Working simultaneously on six screens, the team defined commands which GPT then executed chapter by chapter to create the first version of the manuscript
2. At each stage of the process, the content generated by the Large Language Model (LLM) was reviewed by the authors, who then asked the machine to adapt the text
3. This “prompt ping pong” ensured that the knowledge expertise of the authors renowned in their field was combined with the language expertise of the LLM
4. After the Hack Day, the authors and Springer Nature’s editorial team further checked, corrected and supplemented the text
5. The team then linked the relevant data sources to ensure proper attribution

Henning Schoenenberger, Vice President Content Innovation at Springer Nature, commented: “Given the challenges humanity faces today, as a scientific publisher we are always looking for new ways to accelerate the dissemination of knowledge and thus the development of new insights. Rapidly developing technologies such as Artificial Intelligence offer enormous potential here, but there are also risks. Our job is to ensure responsible use of these tools in the interests of science. This is only possible through human oversight of technology.”

Quick and easy access to trusted science is key to accelerating discovery, as demonstrated by the search for vaccines during the pandemic. AI has the potential to accelerate this yet further. Springer Nature builds on this experience by using technology and AI tools, all of which it is deploying  alongside human intervention and with human oversight, to further help the research community. In 2019, the company published the world's first machine-generated research book. Springer Nature also trials AI to translate or summarise scientific publications, to match research papers with suitable peer reviewers and to detect plagiarism.