At the latest since the hype around ChatGPT and the possibilities that applications based on artificial intelligence are already creating in our lives, AI is the talk of the town again. Prof. Dr Damian Borth and Konstantin Schürholt dedicated their project to the question of how we can make neural networks more secure, trustworthy and sustainable? By analysing populations of neural networks, the researchers want to find out whether there are any common patterns or structures. This would then make it possible to develop approaches and methods to train neural networks even more efficiently. This is a significant issue in relation to sustainability, but also in terms of the security and trustworthiness of the artificial instances.