Crossplag — Cross-Lingual Plagiarism and AI-Text Detection, from a Kosovo Startup to European Assessment Infrastructure

Crossplag was founded in 2016 in Prishtina, Kosovo, by Besart Kunushevci to address cross-language ('translate-to-evade') plagiarism, where a source text is translated into another language specifically to evade Western, largely English-only, detection engines. The platform grew to support around 100 languages and describes an index of roughly 100 billion texts. By the early 2020s it reported more than 5,000 educators and institutions using it worldwide, including individual users at Harvard, Yale and Georgetown.

The tool has been included in independent academic evaluations of plagiarism and AI-text detectors: the European Network for Academic Integrity (ENAI) tested it in 2018 and again in 2020, and it was one of 14 tools independently benchmarked by Weber-Wulff et al. (2023) in the International Journal for Educational Integrity. In July 2023, Crossplag was acquired by the Norwegian company Inspera (terms undisclosed) and folded into Inspera's assessment and remote-proctoring suite, which operates in roughly 160 countries.

ENAI's testing found Crossplag performed well on single-source ('copy-paste') detection but weaker on paraphrase and multi-source detection. The Weber-Wulff et al. (2023) benchmark found AI-text detectors as a category to be 'neither accurate nor reliable' for high-stakes decisions, a caution that applies to Crossplag's AI-content-detection feature as much as to its peers.

Crossplag is now a Kosovo-founded product operating as part of a larger Norwegian company, rather than an independent Kosovar business. Marketing claims about specific accuracy percentages and a named 2025 pilot circulating on third-party AI-tool comparison blogs could not be independently verified against any primary source and are excluded here.