Scientific publishers are facing a growing challenge as generative AI makes it easier to create, manipulate, and duplicate images used in research. Straive and Imagetwin are responding with a new partnership that brings image integrity screening directly into the editorial workflow.
Announced September 23, the partnership integrates Imagetwin’s image analysis capabilities into Straive’s aiKira Editorial Desk API. The integration allows publishers to screen scientific submissions for manipulated, duplicated, plagiarized, or AI-generated visuals alongside more than 20 other automated integrity checks.
The broader aiKira screening process also covers areas including authorship, journal scope, reference verification, and text, giving publishers a wider set of automated checks before a submission advances through the editorial process.
In a pilot study, Straive said the combined solution identified 98% of image integrity issues, including known manipulated images and AI-generated figures. The announcement does not provide details about the size or methodology of the pilot.
Moving image integrity checks upstream
The companies are positioning the integration around a timing problem in scientific publishing. When manipulated or synthetic visuals are identified after publication, publishers may have to investigate and correct the record, potentially creating additional costs and reputational consequences.
By incorporating image screening into the editorial workflow, the companies aim to identify potential issues before manuscripts reach peer review.
“Research integrity is at a breaking point, and this partnership directly answers this ever-growing challenge. Together, we’re empowering publishers to not only uphold the credibility of the scientific record at scale, but get ahead of fraud thanks to the technological breakthroughs we’re achieving,” said Srinivasan Govindarajan, Business Head, Science & Research at Straive.
Imagetwin says its software detects image duplication, manipulation, plagiarism, and AI-generated content, while its database contains more than 160 million academic images.
“The earlier an integrity issue is found, the easier and less damaging it is to fix. Partnering with Straive lets us put image integrity screening directly into publishers’ existing workflows through aiKira, so integrity checks happen automatically, at scale, before problems reach peer review,” said Patrick Starke, Co-Founder and CEO of Imagetwin.
The integration adds image analysis to an existing editorial screening workflow rather than positioning it as a standalone forensic tool. For publishers, the companies argue, that could allow potential integrity issues to be identified earlier, before a paper moves further through the publication process.
16 c
20 c