What to Make of ChatGPT in Scholarly Communications?
/by KnowledgeWorks Global Ltd.
No matter what you are reading, watching, or listening to, everyone today is talking about ChatGPT and how it can help our daily lives and, perhaps, even our work. Released in November 2022 by OpenAI, ChatGPT is an Artificial Intelligence chat bot that “interacts in a conversational way” and it can “answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.”
ChatGPT was quickly embraced by the world for fun and frivolous purposes, but it has also been used to write press releases, news articles, corporate documents, and even essays and journal articles. Though many industries are jumping to see how this new AI can help their business, there are just as many who are being more cautious. In scholarly publishing, there has been much debate about how this new AI can help our industry. KGL has pulled together a round-up of some of the most recent articles to help you make your own decision about ChatGPT.
What IS ChatGPT and how does it work?
On The Scholarly Kitchen, Chef David Crotty breaks down how ChatGPT and other Large Language Models (LLMs) work and how they fit into the larger discussion about AI in scholarly publishing.
An interview with ChatGPT about AI and scholarly communications
In January, Todd A. Carpenter, Executive Director of the National Information Standards Organization (NISO), conducted an interview with ChatGPT about various areas of scholarly publishing – open access, peer review, and even AI itself, which he shared on The Scholarly Kitchen. What he found was that ChatGPT more than passed the Turing Test, though it cannot detect whether or not the answers it gives are correct, just information it has found and produces.
ChatGPT as co-author on journal articles
As soon as ChatGPT went live, journals were starting to see the chatbot listed as an author on papers. In this article in The Guardian, they illustrate the response from these journals, including an updated editorial policy. “Given the frenzy that has built up around this, it’s a good idea to make it absolutely explicit that we will not permit ChatGPT to be an author or to have its text used in papers,” said Holden Thorp, the editor-in-chief of Science.
A look inside the authoring of an article by GPT-3
Scientific American published a look inside what working with GPT-3 (the next version of ChatGPT) is like, when researchers asked it to write a 500 word article about itself and why it could change the publishing process in the future.
ChatGPT and subverting peer review
In this piece in Chemisty World, senior US correspondent Rebecca Trager outlines the challenges with ChatGPT co-authoring papers but also delves into an issue that Andrew White, a chemical engineering and chemistry professor at the University of Rochester experienced. “Just received a 5 sentence very non-specific peer review on a paper. ChatGPT detector puts it at ‘possibly written by AI.’ It reads to me exactly like an AI generated review,” White posted on Twitter. Could ChatGPT subvert the peer review process?
The future with GPT-4
In March, OpenAI released the new version of ChatGPT, GPT-4, and in this article on CNN.com, you can see how the system has improved and how that might impact your work.
Though artificial intelligence has been an area of discussion in the scholarly publishing community for over a decade, the widespread use of ChatGPT in all areas of the publishing process and how that might impact research and the spread of misinformation is likely something that will prompt much more debate and possibly even more stricter rules of usage going forward.
KnowledgeWorks Global Ltd. (KGL) is the industry leader in editorial, production, online hosting, and transformative services for every stage of the content lifecycle. We are your source for peer review services, market analysis, intelligent automation, digital delivery, and more. Email us at info@kwglobal.com.