Dec 21, 2021
American Association for Cancer Research (AACR) and other publishers have integrated AI screening to prevent errors and fraud in research papers.
Since January 2021, the American Association for Cancer Research (AACR) has been utilizing artificial intelligence (AI) to enhance the integrity of scientific images in research papers. This AI-driven screening process, implemented after peer review, aims to identify duplicated or manipulated images—whether they are flipped, stretched, or filtered—before publication.
The adoption of AI for image integrity checks is gaining momentum across the academic publishing industry. While many journals have traditionally relied on manual image screening, some are now automating this process using advanced software solutions like Proofig. The goal is to detect and prevent errors or intentional alterations that could undermine scientific credibility.
AACR evaluated various software options before selecting Proofig, a company based in Rehovot, Israel. The organization has expressed satisfaction with the results, highlighting that AI can help identify problematic images and reduce post-publication issues. However, professional editors remain essential in assessing flagged images, as some duplications occur unintentionally due to mistakes in manuscript preparation.
Experts in the field acknowledge the benefits of AI-powered screening but also note concerns, such as the potential for false positives or the risk of researchers developing ways to bypass detection. Additionally, there is growing interest in cross-journal image comparison, a capability that could further enhance research integrity.
Several major publishers, including SAGE Publishing and the American Society for Clinical Investigation, have integrated Proofig’s AI technology into their review processes.
Meanwhile, other publishers are exploring similar tools, with some investing in proprietary AI-based solutions.
As AI-powered image screening becomes more widespread, the publishing industry anticipates that such tools will soon become a standard part of the manuscript review process—much like plagiarism detection software is today.