The Project

FACTS – Integrity of science in the age of AI

How can academic libraries use modern AI methods to systematically identify unethical and fraudulent publishing practices?

The FACTS project – Fraud Analysis and Counterfeit Testing in Scholarly Publishing – addresses this question. The project responds to a measurable rise in fraudulent scientific findings, driven in part by the availability of generative language models, which poses a threat to public trust in science.
The project centers on developing modern mathematical methods, particularly graph-based models, generative language models, and hybrid approaches for analyzing bibliometric data. The goal is to make the mechanisms behind predatory journals, paper mills, and citation cartels transparent, and to derive evidence-based decision support tools with concrete action options for publication advisory services in Berlin libraries. Additionally, interactive formats will be developed to raise awareness of fake science and foster networking among stakeholders at research institutions.

The project is funded as part of the program “Strengthening Innovation Capacities in Information Provision (STiIV)” from 01.11.23 to 31.10.2026.

Organization

The project is carried out by the Cooperative Library Network Berlin-Brandenburg (KOBV),  which connects the library landscape of Berlin and Brandenburg and performs important  tasks for the member libraries at a central location. KOBV is the only library network in  Germany that is affiliated with a research institute. The close proximity to science and  research makes it possible to develop and test groundbreaking procedures in information  management and to ensure continuous innovation. The KOBV benefits significantly from the  Zuse Institute Berlin (ZIB)’s orientation as a research institute for applied mathematics and  data-intensive high-performance computing.