Aktuell

• von Juliana Fischer
• 10.09.2026
• English version

A video looks authentic, its message sounds plausible – but behind it lies deliberate AI-powered manipulation. Why are some people more susceptible to such content than others? And can we learn to spot manipulation more effectively? These are the questions a team led by Professor Dr German Neubaum at the University of Duisburg-Essen is exploring as part of the new EU-funded COMPASS project. The project has been awarded around €3.5 million through the Horizon Europe programme, with approximately €443,000 going towards psychological research in Duisburg.

Artificial intelligence can generate increasingly convincing images, voices and videos. What may still be exposed by minor inconsistencies today could soon become almost indistinguishable from authentic content. At the same time, the potential for deliberately using such content to deceive people, influence public debate or sow doubt is growing.

COMPASS* therefore looks beyond individual pieces of false information to examine the wider system behind them: How is information manipulation created and spread? Why is it effective, and what can be done to counter it? This also raises a fundamental question: How can democratic societies protect themselves without restricting freedom of expression or privacy? Eight partners from eight European countries are bringing together psychological, legal and societal perspectives to find answers.

At the University of Duisburg-Essen (UDE), Junior Professor German Neubaum leads the psychological view and corresponding working package, entitled “Co-creating Democratic Resilience: from Evidence to Empowerment”. In a series of experiments, the UDE team exposes participants to different types of deepfakes and investigates which factors best predict whether they believe the artificially generated content to be genuine. The researchers consider personal characteristics such as age and education, as well as cognitive abilities, attitudes and emotional factors.

“Why some people are more likely than others to fall for AI-generated misinformation probably cannot be explained by any single characteristic. We expect a combination of different factors, and our experiments will explore how these interact,” says Neubaum, Head of the research group “Psychological Processes of Education in Social Media” at UDE.

The team is also investigating how people’s beliefs shape the way they respond to questionable content. For example, people who consider a message to be socially important may be more inclined to share it, even if they have doubts about its authenticity.

Spotting manipulation before being misled

The next step focuses on prevention. The UDE team is developing and testing what are known as inoculation and prebunking approaches. Participants are familiarised with common deepfake strategies and suspicious cues before the researchers assess whether this prior knowledge helps them identify new forms of manipulation more effectively.

“Such approaches are often compared to a psychological vaccine. But they cannot make people completely immune to disinformation. We want to find out which cues actually help people recognise manipulation – and whether these effects persist when they encounter new content and over longer periods of time,” explains Dr Jana Dreston from the research group “Psychological Processes of Education in Social Media”.

COMPASS stands for “COuntering Manipulation while Preserving Authentic Speech and Society” and will run until August 2029. The interdisciplinary consortium is coordinated by Tallinn University in Estonia. UDE will receive around €443,000 in funding.

Further Informations:
https://cordis.europa.eu/project/id/101287085

Prof. Dr. German Neubaum, Psychological Processes of Education in Social Media, Tel. 0203/37 9-3553, german.neubaum@uni-due.de

Editor: Juliana Fischer, Tel. 0203/ 37 91488, juliana.fischer@uni-due.de