Artificial intelligence is changing science – but not only for the better. Fake quotes, invented studies, and AI-generated specialist articles are appearing more and more frequently.
By Jan Nintemann and Jochen Siegle; Photo by Brano Unsplash
Researchers such as Diomidis Spinellis (Athens University of Economics and Business) and Luis A. Nunes Amaral (Northwestern University) are now warning that the flood of fake studies is growing rapidly and undermining trust in science.
Trust in Danger
When AI invents studies and produces false quotes, the foundation of science suffers: trust in previous findings. This not only leads to reputational damage for researchers, but also threatens scientific progress as a whole.
Professor Luís A. Nunes Amaral from Northwestern University in the USA warns urgently: “If these trends are not stopped, science is going to be destroyed.” Liudmila Zavolokina (University of Lausanne) fears an unstoppable flood of misinformation, loss of trust, and legal risks.
AI Work Falsifies Research into the Opposite
The Australian scientist Will Kitchen (University of Chichester) describes how AI falsified his work – sometimes into the opposite.
A particularly drastic example of fake scientific abstracts: A 2023 study showed that ChatGPT could create completely invented abstracts with fictitious data – around a third of which were so convincing that scientists did not immediately recognize them as fakes.
In another case, researchers demonstrated that AI can generate fake data sets that appear to be plausible clinical data to support a scientific thesis – however, these data sets turned out to be completely invented upon closer inspection.
What is Needed Now
Specialist journals and universities must develop new control mechanisms to expose AI fakes. Springer Nature has already had to withdraw hundreds of AI papers – a warning signal for the entire scientific community.
Discussion at TransformIT Europe 2026
We also take this topic seriously: At the TransformIT Europe Expo & Conference 2026 in Brussels, we will discuss the opportunities and risks of AI in science. And of course, there will once again be the Google Developer Group Conference on the Future of European AI – with leading voices from research, business and politics.
- https://www.springernature.com/de
- https://group.springernature.com/de/group/media/press-releases/uebergabe-ki-tool-an-verlagsbranche/27775178
- https://zavolokina.com/
- https://www2.dmst.aueb.gr/dds/
- https://staff.aub.ac.uk/en/profile/william-kitchen
- European Google Developer Group Conferenceon the future of European AI
- https://news.northwestern.edu/stories/2023/01/chatgpt-writes-convincing-fake-scientific-abstracts-that-fool-reviewers-in-study/
- http://nytimes.com/2025/08/04/science/04hs-science-papers-fraud-research-paper-mills.html