Return to website

Challenges and Strategies in Detection of AI-generated Content

Prof. Dr. Martin Steinebach

Recently, machine learning has led to an increased questioning of the authenticity of images, videos, sounds, and text. Deepfakes make people appear in videos they had no part in. Texts are generated by chatGPT based on a few keywords and lead to problems in teaching. Images can be altered more easily than ever before. There are many approaches to counter these developments. They can either be forensic and examine the content to see if it was created synthetically. Or they try to introduce a traceable hint at the creation stage.  Our talk will provide an overview of content creation methods and countermeasures.