Springer Nature unveils two new AI tools, Geppetto and SnappShot, designed to enhance research integrity amidst rising instances of fraudulent research in the academic publishing community. 

These tools, developed with the Slimmer AI Science division, which Springer Nature acquired in 2023, target two primary indicators of research integrity issues: AI-generated fake content and problematic images.

Addressing the Growing Threat of Fraudulent Research

The academic publishing community has been increasingly affected by fraudulent research, often perpetrated by paper mills or bad actors submitting fake papers containing fabricated data. These deceptive practices undermine trust in scientific research and consume significant time and resources to identify and rectify fraudulent submissions. Springer Nature’s new AI tools aim to detect and prevent such issues preemptively, ensuring the integrity and reliability of published research.

Geppetto: Detecting AI-Generated Content

Geppetto, one of the new AI tools, is specifically designed to detect AI-generated content, a classic indicator of paper mill activity. The tool divides a manuscript into sections and uses its algorithms to assess the consistency of the text in each section. Each section is then scored based on the probability that the text has been AI-generated. Higher scores indicate a greater likelihood of issues, prompting a human check by Springer Nature staff.

Since its implementation, Geppetto has been instrumental in identifying hundreds of fake papers shortly after submission, preventing them from being published and conserving the valuable time of editors and peer reviewers. This proactive approach not only enhances the integrity of the publication process but also deters future fraudulent attempts by showcasing Springer Nature’s robust detection capabilities.

SnappShot: Ensuring Image Integrity

The second tool, SnappShot, is an AI-assisted image integrity analysis tool developed in-house by Springer Nature. It analyzes PDF files containing gel and blot images, looking for duplications—a known integrity problem within the industry. The tool will expand its capabilities to cover additional image types and integrity problems, further speeding up the checks on submitted papers.

Like Geppetto, SnappShot has yet to decide whether a submission should progress through the editorial process. Instead, it signals the need for human assessment of the images, thereby avoiding duplicated or manipulated data publication. This tool also supports authors by providing feedback on their figures, especially in accidental duplications, thus promoting higher research practice and data management standards.

Commitment to Research Integrity

Chris Graf, Director of Research Integrity at Springer Nature, emphasized the importance of these tools in combating the threats posed by fraudulent research. “The publishing industry faces a determined, malicious threat from paper mills or bad actors who submit fake papers containing fabricated data, and we are determined that this content will not get through our systems. These attempts can have serious consequences for trust in science, and investigating and resolving them can take up an enormous amount of time and resources.”

Graf highlighted that the continuous improvement of these AI tools helps Springer Nature stay ahead of fraudsters, ensuring that published research is robust and trustworthy. This commitment is part of a broader effort by Springer Nature to maintain the highest standards of research integrity, which includes significant investment in a growing expert team and extensive technology development.

Collaborative Efforts and Future Developments

In addition to developing these tools, Springer Nature is actively involved in collaborative work with the broader publishing community. As a contributing organization in the STM Integrity Hub, Springer Nature facilitates knowledge and data exchange and participates in developing shared technology tools. This collective effort aims to strengthen the overall integrity of academic publishing.

Springer Nature’s proactive approach, leveraging advanced AI tools like Geppetto and SnappShot, represents a significant step forward in ensuring the credibility of scientific research. By addressing the challenges posed by fraudulent submissions, these tools help maintain trust in academic publishing, supporting the continued advancement of science based on reliable and authentic research.

Conclusion

Springer Nature’s launch of Geppetto and SnappShot reflects a strong commitment to research integrity. These AI tools not only detect and prevent fraudulent research but also support the publishing community in upholding high standards of accuracy and transparency. As Springer Nature continues to refine these technologies and collaborate with industry partners, it sets a robust example of innovation and dedication in the fight against scientific misconduct.

Source: Springer Nature

Image source: COPILOT