Not being educated in anything to do with AI or machine learning, it seems to me that "amplification algorithms" is exactly what these systems are. When I look

Not being educated in anything to do with AI or machine learning, it seems to me that "amplification algorithms" is exactly what these systems are. When I look at a DeepDream image, it appears as though it has amplified many defining characteristics of the original images. Conceptually, reinforcement learning seems very similar to biological positive feedback mechanisms. These mechanisms have guided human development, and cognitive bias is an innate aspect of being human so it stands to reasons that non-biological systems developed in the same way would exhibit biases.