AI Can Now Spot Duplicated Images in the Manuscripts of Journals


Usually, the filtering and checking process of journals before appearing on American Association for Cancer Research (AACR) undergoes unusual extra checks for ensuring credibility. Since January 2021, the AACR has been using Artificial Intelligence or AI software for provisionally accepting the manuscripts before peer reviewing. The aim is to automatically alert editors to duplicate images, including those in which have been filtered, flipped, stretched, and rotated. 

The AACR is an early adopter of the AI, to avoid pushing papers with images that have been doctored whether because of outright fraud or inappropriate attempts to beautify findings many journals have hired people to manually scan the manuscripts to help check what they find. But few of the publications have already started automating the process by depending on AI software to spot partial and full duplications before manuscripts are published. 

The AACR tried numerous software products before it settled on a service from Proofing, a firm in Rehovot, Israel. Many professional editors are still needed to decide what to do when the software flags images. If data sets are deliberately shown twice with explanations then repeated images might be appropriate for instance. Now, AI is getting sufficiently effective and low-cost. Publishing-industry groups also say they are exploring ways to compare images in manuscripts across journals. 

Researchers have been developing image-checking AI for years due to their concerns over the frauds which may pollute the scientific literature to a much greater extent than the limited numbers of retractions and corrections suggest. Proofig’s software extracts images from papers and compares them in pairs to find common features, including partial duplications. The cost of image checking is much greater than that of plagiarism checking, which specialists say can run to less than US$1 per paper. 

Publishers who have not adopted AI image screening cite cost and reliability concerns are working on their own AIs. Elesevier says it is still testing the software although it notes that some of its journals screen all accepted papers before publication, checking for concerns around images using a combination of software tools and manual analysis. 

Pulverer says that EMBO Press still mostly uses manual screening because he’s not yet convinced by the cost-benefit ratio of the commercial offerings, and because he is a part of the cross-publisher working group that is still defining criteria for software. 

AI image checking is generally done within a manuscript, not across many papers which would make it increasingly computationally intensive. But the thing that academic and commercial software developers say is that it is technically feasible. 

Crossref, a US-based non-profit collaboration of more than 15,000 organizations that organizes plagiarism checking across papers, among other things, is currently running a survey to ask its members about their concerns on doctored images, what software they are using and whether a “cross-publisher service” that could share images could be viable and helpful, says Bryan Vickery, Crossref’s director of product in London.