Thousands of scientific journal publishers have prohibited or restricted contributors’ use of a sophisticated AI-driven chatbot, fearing that it may contaminate academic literature with inaccurate and even fake research.
Springer Nature, the world’s largest academic publisher, is one of them, having defined its standards on the use of AI writing tools in scientific publications. This week, the company announced that software such as ChatGPT cannot be attributed as an author in papers published in its thousands of publications.
“We felt compelled to clarify our position: for our authors, our editors, and ourselves,” Magdalena Skipper, editor-in-chief of Springer Nature’s flagship publication, Nature, told The Verge.
“This new generation of LLM tools — including ChatGPT — has really exploded into the community, which is rightly excited and playing with them, but [also] using them in ways that go beyond how they can genuinely be used at present.”
However, Springer claims it has no problem with scientists utilizing AI to help them write or generate ideas for study as long as the authors correctly declare their involvement.
“Our policy is quite clear on this: we don’t prohibit their use as a tool in writing a paper,” Skipper tells the Verge. “What’s fundamental is that there is clarity. About how a paper is put together and what [software] is used.”
“We need transparency,” she added, “as that lies at the very heart of how science should be done and communicated.”
ChatGPT and previous large language models (LLMs) have already been identified as authors in many papers, preprints, and scholarly journals. However, the type and extent of these tools’ contributions differ by the situation.
The scientific community’s reaction to works citing ChatGPT as an author has been primarily negative, with social media users labelling it “absurd,” “silly,” and “very stupid.”
As Skipper and Springer Nature explain, one of the arguments against assigning AI authorship is that software simply cannot fulfil the needed tasks.
The software cannot be held legitimately accountable for a publication, cannot protect its work’s intellectual property rights, and cannot interact with other scientists or the media to address questions or issues about its work.
However, there are several compelling reasons to use ChatGPT in the field. Skipper argues that academics whose first language is not English can benefit from AI writing tools.
“It may be a levelling tool from that perspective,” she says.
“I think we can safely say,” Skipper continued, “that outright bans of anything don’t work.”