Science & technology

Study: Over Half of Peer Reviewers Now Use AI to Evaluate Research

Does this enable manipulation and undermine integrity, or does it responsibly enhance efficiency and rigor?

Published DEC 16 2025

Study: Over Half of Peer Reviewers Now Use AI to Evaluate Research

Story

Above: A phone screen displaying an AI sign, along with a silhouette of a paper shaped like a human face and binary code, on Jan. 15, 2023. Image credit: Jakub Porzycki/Contributor/NurPhoto via Getty Images

The Facts

A survey conducted by Frontiers involving 1,645 active researchers across 111 countries found that 53% of peer reviewers now use artificial intelligence tools when evaluating manuscripts, with nearly one in four respondents reporting increased AI usage over the past year.
Nature
Study Finds
PHYS
Among reviewers who use AI in peer review, 59% deploy it for drafting reports or improving grammar, 29% use it to summarize manuscripts or identify gaps, 28% use it to flag potential misconduct such as plagiarism and 19% use it to assess methodology or statistical soundness.
Study Finds
Nature
Early-career researchers show the highest AI adoption rates at 87%, compared to 67% among senior researchers, while regional adoption varies with 77% in China, 66% in Africa, 46% in Europe and 31% in North America, according to the Frontiers whitepaper findings.
Study Finds
Frontiers
While 66% of researchers surveyed rate publisher use of AI as effective for speeding up publication, only 21% say it increases their trust in the process, and 76% report being unsure or unaware whether publishers used AI during their publication experiences.
Entrevue.Fr
Study Finds
A separate survey by IOP Publishing found that 41% of respondents now see benefits of AI in peer review, up from 12% the previous year, while 37% hold negative opinions, leaving 22% neutral or unsure.
Study Finds
Times Higher Education (THE)
Frontiers has set out policy recommendations including mandating transparency around AI use, embedding AI literacy training across the research system, strengthening integrity standards, improving data provenance and auditability, and ensuring equitable access to trustworthy AI tools.
PHYS
Frontiers
Research Information

Sources Split

LEFT

CENTER

RIGHT

ESTAB
CRIT

CENTER

PRO
ESTAB

The Spin

Techno-optimist narrative

AI adoption in peer review represents responsible innovation that addresses the overwhelming volume of research publications modern reviewers face. With proper transparency requirements, training and governance frameworks, AI tools enhance efficiency and clarity while supporting methodological rigor. The research community shows strong enthusiasm for AI-assisted review, particularly among early-career researchers and in rapidly growing research regions.

Frontiers

Research Information

Nature

Techno-skeptic narrative

Hidden AI prompts in academic papers expose a crisis in peer review integrity, with researchers deliberately manipulating AI-powered reviewers through invisible text instructions. These prompt injections represent a sophisticated form of peer review manipulation that undermines the entire quality control system of academic publishing. The practice reveals how widespread AI use by reviewers has created new opportunities for misconduct that current research guidelines fail to address.

The Conversation

The Japan Times