PhD Position in Explainable DeepFake Detection Using Foundation Models

This PhD position at the University of Twente focuses on advancing explainable AI techniques to detect and analyze DeepFakes—manipulated videos and audio—using foundation models. The project is part of the EU-funded DEFORM initiative, which brings together 14 research institutes and companies across six countries to create tools for reliable, interpretable detection of synthetic media. The successful candidate will collaborate closely with international partners to explore inconsistencies in video and audio data, ensuring both detection and explanation of manipulation traces. The role emphasizes interdisciplinary collaboration, hands-on technical work, and a strong foundation in computer vision and AI. The position is fully funded for four years, offering a dynamic research environment with opportunities for international travel and professional development.

• Develop and implement AI methods to detect manipulated videos and audio using foundation models and other advanced techniques.
• Analyze inconsistencies in video and audio data to explain traces of manipulation and identify synthesis tools used.
• Collaborate closely with international partners in the EU DEFORM project to integrate findings into a unified detection framework.
• Conduct experimental validation of methods, including testing under real-world conditions and uncertainty analysis.
• Present research findings at project meetings, conferences, and workshops, both nationally and internationally.
• Contribute to the development of tools and methodologies for explainable detection of synthetic media.
• Work in an interdisciplinary environment, bridging gaps between computer science, engineering, and forensic science.