MDPI Adopts AI to Uphold Scientific Image Integrity

In a significant move to bolster scientific credibility, open access publisher MDPI has announced a multi-year partnership with Proofig AI, a company specializing in AI-driven integrity software. This collaboration aims to systematically check for image manipulation in academic papers, reinforcing the standards of research integrity.

A Strategic Partnership to Enhance Research Integrity

Following a successful pilot program, MDPI will begin implementing Proofig AI’s image analysis tools across a selection of its biomedical journals. The initial rollout is the first step in a broader plan to extend these checks to all relevant submissions. The pilot demonstrated the system's effectiveness, improving the detection of problematic figures and streamlining the editorial review process.

Dr. Dror Kolodkin-Gal, co-founder and CEO of Proofig AI, expressed his enthusiasm for the collaboration: “By combining Proofig AI’s advanced technology with MDPI’s editorial expertise, we can help prevent errors and maintain integrity in science, while setting a new benchmark for research integrity and quality control in publishing.”

How AI is Revolutionizing Image Verification

Proofig AI's software is designed to tackle the growing challenge of image manipulation, whether accidental or intentional. The system automatically flags images for editorial review by checking for common issues such as:

• Alteration and manipulation
• Duplication and reuse of images
• Plagiarism from sources like PubMed
• Undisclosed AI-generated images

It also provides specialized tools for analyzing western blot images, a task that is often time-consuming and complex for human reviewers. During the pilot, the software successfully identified issues like duplicated histology images, including instances where figures were altered to hide their similarity.

Tim Tait-Jamieson, Head of Publication Ethics at MDPI, noted the importance of this technology. “The life sciences are disproportionately affected by research integrity issues, making vigilance in this field especially critical,” he said. “By integrating Proofig AI into our editorial workflows, we strengthen our ability to detect scientific misconduct early and reduce post-publication amendments.”

Proven Success and Future Rollout

MDPI conducted a comparative evaluation of image detection tools for over a year, and Proofig AI emerged as the clear choice due to its quality, ease of use, and positive feedback from internal teams. The initial results of the integration are already promising, with MDPI reporting a year-on-year decrease in post-publication flags related to image manipulation.

“We’re thrilled to be moving forward with this partnership, ensuring we do our utmost to protect MDPI’s biomedical journals,” stated Sanita Meijere, IT Product and Project Manager at MDPI. As the new workflows stabilize, the publisher plans to expand the use of Proofig AI's technology beyond the initial set of journals, reaffirming its commitment to maintaining the highest standards in scientific publishing.