Springer Nature’s Research Integrity Group will use the tool, produced in-house using AI, to assess submissions to most of the journals and books it publishes, analysing the relevance of each reference used.

It is the latest tool Springer Nature has built as it looks to “to weed out problematic submissions and ensure the veracity of the publication record”, the company said.

If the tool identifies several irrelevant references, it will flag the submission to the integrity team to manually check the manuscript and decide whether it should be withdrawn.

Springer Nature is the publisher of brands including scientific journal Nature and trade publishing business Palgrave Macmillan.

Chris Graf, director of research integrity at Springer Nature, said: “As the use of AI to generate fake research papers becomes more effective, reference checking provides a key opportunity to identify these unethical efforts. 

“The addition of this tool to our existing checks will help us to catch many unethical submissions that would otherwise progress to take up editors’ and peer reviewers’ time.”

Last year, as well as forming the 50-person research integrity unit, the company launched two other tools to tackle the problem – one helping to identify papers containing AI-generated fake content and another flagging submissions with ‘problematic’ images.