Journal editors, researchers, and publishers are debating the role of AI tools like ChatGPT in published literature and whether they should be cited as authors
Chris Stokel-Walker is a freelance journalist in Newcastle, UK. — You can also search for this author in — PubMed Google Scholar Tweets: @drkahaynes , @neuroecology , @jazliaziz , @anantbhan , @agbioworld , and @glynmoody Tweets: Dr. Karmella Haynes, Ph.D. / @drkahaynes : Hot take: ChatGPT is a sophisticated calculator; a tool like any software. A person who sees it as a “co-author” doesn't understand what co-authorship means. As the article notes, a co-A must be able to take responsibility for the research findings. https://www.nature.com/... Adam J Calhoun / @neuroecology : Scientists to ChatGPT: 😤 https://www.nature.com/... Jazli Aziz, PhD / @jazliaziz : We have now reached the point where artificial intelligence has been listed as a co-author on scientific publications. Clearly this is illogical as the bot merely scraps data from the internet and doesn't have any “intelligence” of its own. https://www.nature.com/... Anant Bhan / @anantbhan : Soon ChatGPT will have the highest h index in the world? ChatGPT listed as author on research papers: many scientists disapprove At least four articles credit the AI tool as a co-author, as publishers scramble to regulate its use. https://www.nature.com/... Channa Prakash / @agbioworld : The AI chatbot #ChatGPT that has taken the world by storm has made its formal debut in the scientific literature — racking up at least four authorship credits on published papers and preprints. Journal editors scratching their heads. @OpenAI https://www.nature.com/... Glyn Moody / @glynmoody : ChatGPT listed as author on research papers: many scientists disapprove - https://www.nature.com/... no end of interesting questions raised by #AI...
Context & Ripple Effects
The authorship question has already been answered in practice before the debate caught up: at least four papers and preprints have credited ChatGPT as a co-author, drawing disapproval from scientists who argue co-authorship requires taking responsibility for findings — a point Karmella Haynes makes bluntly in the article's quoted reactions. Publishers are now scrambling to codify what practice has muddied.
Springer Nature set the template early with a policy barring LLMs from authorship while allowing disclosed AI assistance, and the surrounding coverage explains why editors are cautious: researchers report chatbots producing convincing errors on technical topics (apprehension documented by Nature), and separate work shows ChatGPT can launder conspiracy theories into clean prose (per NYT-sourced research). The stakes are the integrity of the authorship record itself.
First-order effects
- Journals and publishers must draft or clarify AI policies immediately, since at least four papers have already credited ChatGPT as co-author and each new submission forces an editorial ruling.
Second-order effects
- Researchers using ChatGPT for drafting face a disclosure regime rather than a ban at major publishers like Springer Nature, shifting the compliance burden to authors while giving reviewers a new failure mode to check for undisclosed generation.
Third-order effects
- If the Springer Nature model spreads, academic credit systems will harden around human accountability — AI becomes acknowledged tooling like software, never an accountable party — forcing every publisher to define 'disclosed use' or risk an authorship record that can't assign responsibility.
The trend: Academic publishing is converging on a tool-not-author settlement for generative AI, with disclosure rules replacing authorship credits as the enforcement mechanism.