Remarkable Stories: AI & Technology Innovations Across EUTOPIA

Artificial Intelligence (AI) continues to shape our world—and EUTOPIA universities are right at the centre of this transformation. In this month’s edition of Remarkable Stories, we spotlight three standout projects showing how AI and machine learning are being used to solve real-world problems: from tackling deepfakes and disinformation to improving urban safety and advancing cancer research.

University of Ljubljana: Advancing Deepfake Detection

In an era where digital forgeries threaten the integrity of information, researchers at the University of Ljubljana have built a new defence. A team led by Dr. Borut Batagelj and Dr. Peter Peer from the Faculty of Computer and Information Science—working closely with colleagues from the Faculty of Electrical Engineering—has developed an AI model called M-Task-SS, designed to detect and localise deepfakes with high precision.

Unlike many existing tools, M-Task-SS is trained only on real images. It generates its own forgeries to improve detection accuracy, combining classic and transformer-based neural networks. The result? A model that doesn’t simply flag an image as fake—it pinpoints where exactly the image has been manipulated, offering a clear, visual explanation. This transparency is invaluable to forensic analysts, journalists, and digital platforms.

Most impressively, the model performs better than its peers—especially against forgeries made with diffusion models, one of the newest and most powerful deepfake techniques. It’s a leap forward for media verification and digital trust.

But it's not just researchers taking on this challenge at EUTOPIA. In May 2025, students from across the alliance gathered in Barcelona for the Innovation Challenges for Students at Pompeu Fabra University, under the theme “Unmasking the Truth: Tackling Fake News Together.” Over three days, students collaborated on solutions to the growing crisis of disinformation, from deepfakes and mental health impacts to political manipulation and fact-checking systems. 

Andrés González-Nandín, journalism professor at UPF and director of Diari de Barcelona, stressed the urgency of the challenge:

“Disinformation and fake news are issues that institutions all over Europe are very concerned about. They pose a real challenge regarding how to critique and inform society on decisions that must be made with reliable information… Some solutions proposed by students and researchers are imaginative and could be implemented fast with the right funding and partnerships—from fact-checking organisations to EU institutions.”

Together, this cutting-edge research from the University of Ljubljana and the student-led innovation displayed in Barcelona capture the EUTOPIA spirit: achieving academic excellence while creating real-world impact.

Check out the links to find out more:

• ​​University of Ljubljana - New Tools to Combat Deepfakes
• Innovation Challenges for Students at UPF in May 2025

University of Gothenburg: Predicting Pedestrian Behaviour with Machine Learning

Each year, traffic accidents claim 1.19 million lives globally—more than half of them pedestrians. Improving the prediction of pedestrian behaviour could reduce these numbers significantly. In a recent study, Chi Zhang, a doctoral researcher at the University of Gothenburg, and colleagues are doing just that, through machine learning. Th research team have been able to predict pedestrian's decisions to cross a road with better accuracy than ever before, and more quickly predict pedestrian trajectory. 

Their work focuses on two core questions: will a pedestrian cross the road, and if so, what path will they take? The machine learning models achieved incredible accuracy in predicting crossing decisions and crosswalk usage, demonstrating a big step forward in this field. Zhang explains: 

“In over 90 percent of cases, the study was able to successfully predict whether a pedestrian would cross the road or wait, which is 4 percent higher than existing models. With over 94 percent accuracy, we could successfully predict whether the pedestrian would use a zebra crossing, a result also 4 percent better than existing models.”

Using data from VR simulations in Japan and Germany, the study also examines the psychology behind crossing decisions—how factors like waiting time, walking speed, and missed opportunities shape risk-taking behaviour. For instance, people who miss several crossing “gaps” become more likely to take risks. Zhang also developed a deep learning model using real-world data from the U.S., Switzerland, and Cyprus. It predicts a pedestrian’s path using only a few seconds of movement history, and does so faster than prior systems—crucial for real-time use in autonomous vehicles. Zhang states: 

"This quick data processing has the potential to be used in the development of self-driving cars, where it is crucial for the vehicle to receive data as quickly as possible”

Chi Zhang believes the study’s findings on pedestrian behaviour in traffic are valuable both for the development of autonomous vehicles and for urban traffic planning, whilst utimately playing a significant role in reducing the number of fatal pedestrian accidents . This research is a clear example of how AI and machine learning, when applied responsibly, can improve everyday life and shape safer cities.

Check out the link to find out more:

• University of Gothenburg - Machine learning predicting pedestrian behaviour in traffic

TU Dresden: Deciphering Cancer Metastasis Through AI

At TU Dresden, researchers are turning to AI to crack one of cancer’s biggest mysteries: metastasis. The DECIPHER-M project—short for Deciphering Metastasis with Multimodal Artificial Intelligence Foundation Models—is developing cutting-edge AI tools to understand and predict how cancer spreads.

Led by Prof. Dr. Jakob N. Kather at the Else Kröner Fresenius Center for Digital Health, the project brings together leading experts in medicine, computer science, and biotechnology from institutions across Germany. The approach? Train AI systems to detect patterns in everyday clinical data that are invisible to the human eye.

The models integrate data from multiple sources, forming a “multimodal” foundation that helps clinicians assess metastatic risk early on. This could lead to faster diagnoses, fewer unnecessary treatments, and more personalised care—potentially improving survival rates and quality of life for cancer patients. Project Coordinator Professor Kather explains:

“Despite enormous progress in oncology, metastasis remains one of the biggest challenges in cancer treatment. Our multimodal approach allows us to predict individual risk more precisely and develop personalised treatment strategies.”

Funded with €5.5 million for the first three years by Germany’s Federal Ministry of Education and Research (BMBF), DECIPHER-M is part of the government’s “National Decade Against Cancer” initiative. For TU Dresden’s Faculty of Medicine, this is a prime example of what interdisciplinary research can achieve. Professor Esther Troost, Dean of the Carl Gustav Carus Faculty of Medicine at the TU Dresden adds: 

“The goal is clear: improve the quality of cancer treatment, avoid ineffective therapies, and ease the burden on the healthcare system. In the long term, DECIPHER-M could help reduce cancer mortality and improve patients’ lives.”

As AI reshapes the future of medicine, projects like DECIPHER-M show the power of combining data, technology, and cross-disciplinary expertise to tackle some of the world's most complex health issues.

Check out the link to find out more: 

• TU Dresden: Using Artificial Intelligence to Decipher the Mechanisms of Cancer Metastasis

In a world where the risks of AI often dominate the headlines, these projects highlight a more constructive path—one grounded in responsibility, collaboration, and impact. EUTOPIA’s researchers and students are not just keeping pace, they’re shaping how these technologies are used to meet real-world challenges and improve society in this rapidly evolving landscape.

The topic of artificial intelligence will also be in the spotlight for the upcoming Research Days on the 1-2 July during the EUTOPIA Week in Warwick, where researchers from across disciplines will present their latest work on the opportunities and challenges of AI. Stay tuned for more information!


Written by Nathan Lindstrom