Error message

  • Notice: A non well formed numeric value encountered in _video_filter_process() (line 216 of C:\inetpub\drupal\sites\all\modules\video_filter\video_filter.module).
  • Notice: A non well formed numeric value encountered in _video_filter_process() (line 258 of C:\inetpub\drupal\sites\all\modules\video_filter\video_filter.module).

Report on the Use of Artificial Intelligence by Researchers from Springer Nature

Primary tabs

We invite you to review the mini‑report published by Springer Nature in December 2025, based on a survey of 3,825 researchers worldwide regarding their use of artificial intelligence (AI) in scientific work — Researchers’ Use of AI for Content Discovery and Submission Decisions.

Main Content of the Report

Use of AI for Content Discovery

Prevalence: Approximately 23% of researchers use AI tools to search for scientific content. However, AI is mostly used as an additional tool alongside traditional methods (specialized web search, publisher platforms, library databases).
Purpose of Use: AI is applied for three main tasks:
  1. Finding specific articles or data.
  2. Synthesizing knowledge and generating summaries.
  3. Generating ideas and brainstorming.
Demographics: Early‑career (junior) researchers use AI more often (27%) than senior researchers (22%).

Use of AI for Submission Decisions

Low level of adoption: Only 14% of researchers use AI to decide where to submit their manuscripts.
Traditional factors dominate: When choosing a journal, researchers still rely primarily on impact factor, journal reputation, and prior experience rather than AI recommendations.
Age differences: Early‑career researchers use AI for journal selection twice as often (20%) as senior scholars (10%).

Trust Levels and Regional Differences

Skepticism: Overall trust in AI recommendations remains low. Most researchers either do not trust AI or take a neutral position (“it depends on the circumstances”).
Geographical divide: There is a strong contrast between East and West. Researchers from China and India show significantly higher trust and adoption of AI than those from Europe and North America. For example, almost half of researchers from North America (45%) do not trust AI for journal selection, while in China the distrust level is four times lower.

Popular Tools

Clear leader: The most popular tool remains ChatGPT (used by 83% for content discovery and 76% for submission decisions).
Other tools: The use of Gemini is increasing. Researchers also use specialized tools such as SciSpace, Scite, DeepL, and Elicit.

Role of Libraries

Low involvement: Only 9% of researchers believe their library is “highly involved” in overseeing the use of AI. A quarter of respondents do not know whether their library participates in this process at all.
Regional specifics: Libraries in China are much more actively integrated in AI‑related processes compared to libraries in Western countries.
By the end of 2025, AI remains in an experimental phase of use. It does not replace traditional methods of searching and evaluating scientific information, but rather complements them. A clear divide in the perception of AI persists: early‑career researchers and scholars from Asian countries are far more optimistic about AI than older generations and researchers from Western regions.