Nature Survey Reveals What 5,000 Researchers Really Think About AI Writing

2 min read


TL; DR

A Nature survey of over 5,000 researchers shows that most people accept using AI tools for translation, proofreading, and even writing the first draft—if disclosed. But using AI for peer review is widely frowned upon. Transparency is key.


AI writing tools are changing the way researchers work—faster drafts, better grammar, smoother translations. But some big questions remain:

📍 How much AI is too much?

📍 Is using AI in academic writing ethical?

📍 Should authors always disclose their use of AI?


A recent Nature news feature, titled "Is it OK for AI to write science papers? Nature survey shows researchers are split" (https://www.nature.com/articles/d41586-025-01463-8), provides valuable insights. Based on responses from more than 5,000 researchers across disciplines and career stages, the survey explored how AI is being used and what researchers think about it.


Overwhelming support for editing and translation

More than 90% of respondents said they were comfortable using generative AI to edit or translate their scientific writing. This includes improving grammar, clarity or structure, especially for those who are not native English speakers.


"I don't think it's any different than asking a colleague who is a native English speaker to read and revise your text," one respondent noted.



Two-thirds are OK with drafting—cautiously

Roughly two-thirds of researchers said it was acceptable to use AI to help write manuscripts, such as generating first drafts or summarizing content.


But the support was less enthusiastic than for translation and editing. Many researchers emphasized that AI should assist, not replace, the author’s thinking and expertise.


One respondent put it bluntly: "You can use AI to help you write your paper, but not to think for you."



Most researchers reject AI-written peer reviews

When it comes to peer review, attitudes shift.

More than 50% of respondents disapproved of using AI to write referee reports, citing concerns about confidentiality and scholarly responsibility.



Should you disclose AI use?

This was one of the most debated questions in the survey.

While many felt that basic editing or translation didn't require disclosure, most agreed that using AI to generate content—such as paragraphs or entire drafts—should be disclosed.


Journal policies vary widely:

📍 Lenient: Journals like Springer Nature and IOP Publishing say that light editing doesn't require disclosure, as long as  the authors take responsibility.

📍 Moderate: Wiley suggests that if AI is used to generate or rewrite text, authors should include a note in the manuscript.

📍 Strict: The JAMA journals require authors to specify which AI tools were used, how they were used, and to take full responsibility for any AI-generated content.


"It’s like disclosing a conflict of interest," said one journal editor. "It’s not that we mind if you use AI—we mind if you don’t tell us."



Are Stork's AI Tools Still Safe to Use?

Definitely—and wisely. Stork's AI tools are here to help researchers work more efficiently.

📍For drafting with For Writing Assistant and Translation: They can be incredibly helpful—especially for researchers writing in a second language. In fact, over 40 published papers have used Writing Assistant.

📍 For drafting with AI paper: It can be a great starting point. But you should personally review, verify, and revise the draft AI generates.

📍 Before submission: Always check the journal's AI use policy. When in doubt, disclose AI involvement.



The academic world is still figuring out how best to use AI in research. Surveys like Nature's are just the beginning.

How do you use AI in your writing process? Please let us know.





写作助手,把中式英语变成专业英文


AI writing papers with real references


Want to receive new post notification? 有新文章通知我

Two in a row! Stork helped Xu’s team publish…

    Info "The authors extend their gratitude to Stork's Writing Assistant (https://www.storkapp.me/writeassistant) for its assistance in polishing some of the text in this article"    Recently, Yuehua Xu and he
Wanling Zhu
1 min read

连发两篇!常州大学徐月华团队使用文献鸟发表文章

你知道吗? 如果您发表的文章引用了文献鸟Stork,请告诉我们([email protected])。我们会根据期刊的影响因子给您优惠。 近日常州大学电子与控制工程学院的徐月华团队连发的两篇文章引
Wanling Zhu
21 sec read

ChatPaper 更新:思维导图让文章脉络更清晰

懒人总结 ChatPaper 更新,增加了思维导图,帮助您梳理文章框架,让文献阅读更有效率。您还可编辑和保存思维导图用于汇报展示或笔记记录,让复杂的研究逻辑更容易被讨论和共享。 研究表明,思维导图在概
Wanling Zhu
5 sec read

Leave a Reply

Your email address will not be published. Required fields are marked *