New report: Feasibility of technical solutions for the detection of falsified images in research

Amidst the ongoing issues of research integrity and paper mills, image integrity is a growing concern. The rise of generative AI poses significant additional risks, as it allows for the generation of images and other types of data that are very hard if not impossible to detect with current technologies. For this reason, STM funded a report developed by MoreBrains Cooperative, looking into the feasibility of technical solutions for the detection of falsified images in research.

Organisational Challenges and a Collaborative Solution

The report, which primarily looked at immunohistochemistry blots and microscopy images, explores current approaches to verifying image integrity, challenges and possible solutions.

Through a mix of desk research and interviews with stakeholders, we found technical solutions are available to establish the authenticity of images, but there are significant challenges around standards, governance and community buy-in. To be effective, an image integrity system must be available to and accepted by all.

“We hope that this report provides a clear overview of possible solutions to establish whether images are authentic — an important step in assessing the integrity of submitted research,” quotes Joris van Rossum, Program Director, STM Solutions. “At the same time, it makes clear that the greatest challenges do not lie in this technology, but in the socio-technical aspects of its implementation on a broad scale. During the coming months, this report will form the basis of discussions with several stakeholders in the research ecosystem to assess whether and to what extent the proposed solutions can be implemented.”

Learn More 

Download the full report to learn more about the proposed solutions, which would draw on digital signatures, trusted certificates with distributed registration and a persistent identifier system.

Join us on December 4 for STM Integrity Day in London where this report will be discussed, and the Researcher to Reader conference in February of 2025. STM will follow up with a summary of the outcome of the discussions and a proposed next step in the spring of 2025. 

Got feedback? We’d love to hear your thoughts