Researchers’ AI expectations undergo “significant reality check”

Wiley has announced key findings from its second ExplanAItions study, revealing a “major course correction” taking place.

The study of more than 2,400 researchers worldwide finds that they remain optimistic about AI, with 85% reporting that it has improved their efficiency, and close to three-quarters saying it has enhanced both the quantity and quality of their work. Overall use of AI tools surged from 57% in 2024 to 84% in 2025, including specific use for research and publication tasks, which grew significantly to 62% from 45%. 

However, while AI usage has surged dramatically, researchers are significantly scaling back their expectations of what AI can currently do as they gain firsthand experience, moving beyond hype toward nuanced, evidence-based adoption. Last year, researchers believed AI already outperformed humans for more than half of potential use cases presented. This year, that figure dropped to less than one third.

“We’re witnessing a profound maturation in how researchers approach AI as surging usage has caused them to recalibrate expectations dramatically,” said Jay Flynn, Wiley EVP & General Manager, Research & Learning. “Wiley is committed to giving researchers what they need most right now: clear guidance and purpose-built tools that help them use AI with confidence and impact.”

Increased hands-on experience has bred more informed caution among researchers, with concerns about potential inaccuracies and hallucinations rising significantly from 51% to 64%. Privacy and security concerns have similarly intensified, climbing from 47% to 58%. This evolution reflects a maturing research community moving beyond initial enthusiasm toward a more nuanced understanding of present limits and future potential.

Researchers are more likely to rely on general-purpose AI tools rather than specialised ones for science and research, with 80% using mainstream tools like ChatGPT compared to just 25% using AI research assistants. The root of this disparity lies partially in awareness, with only 11% of researchers on average having heard of the specialised research tools surveyed. Most researchers (70%) are using a freely accessible tool, even if they are among the nearly half (48%) who also have access to a paid solution. Wiley says this suggests that many researchers are employing a patchwork of solutions to try to meet their needs. 

The study reveals a significant disconnect between researcher needs and institutional support, with only 41% feeling that their organisation offers adequate support. To help fill this gap, most researchers (73%) are looking to publishers to provide clear guidelines and help them avoid pitfalls. More than half of researchers (57%) cite lack of guidelines and training as a primary barrier to expanded AI adoption. 

New ExplanAItions data comprises views from 2,430 researchers surveyed in August 2025. This new report builds on the methodology established in the original ExplanAItions study (published in February 2025, with data from April and August 2024). This initiative aims to track perception and use of AI among researchers year-over-year, presenting a global analysis. 

An initial set of findings is available now via the Wiley website, with the full findings to be released at the end of October.