NEWS

Early-career researchers lead shift towards AI-first discovery
in research workflows

Global data released in The State of AI Use Among Researchers, a new white paper from Springer Nature¹, shows that AI is becoming embedded at the earliest stages of the research process, reshaping how researchers find, assess, and prioritise information. Drawing on more than 10,400 qualified global responses it showed that nearly 50% of respondents report frequent AI use for information gathering and editing, highlighting the value AI provides to researchers at the start of their work. AI usage remains lower for tasks such as reviewing papers and submission decision-making, reflecting the continued importance of accountability and professional judgement.

The data also points to a clear career-stage divide. Around 48% of early-career researchers report frequent AI use for information gathering and improving or editing papers compared to around 34% of researchers more established in their careers². This suggests an indication in how research practices may evolve, and the role publishers can play in supporting responsible, effective AI use.

The latest white paper has been produced with Fudan University and the Shanghai Academy of AI for Science to provide greater insights into attitudes towards AI across the global research community. The white paper provides an interesting snapshot of evolving behaviours across the research community, helping to benchmark adoption and track how the role of AI is developing over time.

Other key findings indicated:

• General-purpose tools are most used: General-purpose AI models account for 75.9% of all tool mentions, while specialised writing and editing tools or literature and evidence discovery tools show lower engagement.

• Institutional support is still limited: Only 11% report institutional funding for their primary AI tool, while most use free tools or pay personally.

• Concerns over AI reliability persist. While nearly two-thirds of respondents were satisfied or very satisfied with AI for content discovery, concerns remained over the accuracy and reliability of AI-generated recommendations – reinforcing the importance of trusted content in research workflows and tailored AI tools for research.

Saskia Steinacker, Chief Product Officer at Springer Nature, said: “Understanding how researchers are using AI is critical to ensuring its adoption remains researcher-led. By working with partners and listening to the research community, we can develop tools where they can have the greatest, most effective impact. These findings show AI’s value in discovery and navigating complex information, while reinforcing the importance of trust, transparency and verification.”

“At Springer Nature, we are focused on developing AI solutions that help researchers work more efficiently and improve access to trusted content, while upholding the quality, integrity and responsibility that research depends on.”

This forms part of Springer Nature’s broader approach to understanding and responding to researcher needs. The company continues to report transparently on its AI strategy, investment and development, while providing guidance to the research community through its AI hub. More information can also be found in Springer Nature’s annual report and media centre.

The full white paper The State of AI Use Among Researchers from Nature Research Intelligence, Fudan University and the Shanghai Academy of AI for Science can be found here: https://www.nature.com/immersive/partnercontent/fudan/aiforscience/index.html

1. Global data is from Springer Nature's survey, ‘The State of AI Use Among Researchers’, fielded in March 2026. The survey drew 10,480 qualified responses, spanning researchers from 117 countries or economies across 12 research disciplines.

2. Defined as researchers with fewer than three years’ experience and researchers with 20 or more years’ experience.