In recent years we can observe that with the help of AI and Machine Learning we can generate realistic looking images that confuse humans. DeepFakes, GAN based algorithms, Transformers type of AI architectures give us realistic images of a non-existent reality. In many tasks, humans lose while playing against AI. Lee Se-dol while losing against the machine had no options to do a similar gesture as the procrastinating character played by Kurt Russel did in the opening scene of John Carpenter's "The Thing".

For this semester, our aim is to use Machine Learning to develop a Neural Network algorithm that can generate an output which appears realistic for AI. We're gonna explore the plausible aesthetics of abstract patterns which creates a cognitive illusion for machine perception. Let's say we improve Machine Learning to Machine Confusion to make something which is perceived "deep fakeish" from the point of view of AI.

Keywords: AI, Neural Nets, Python, AudioVision, Machine Perception