on We have grown accustomed to the insane hyperbolic proclamations of AI CEOs. From claiming their technology could wipe out humanity to stating they are close to producing a superintelligent AI, every time they open their mouth, unfounded manipulative PR drivel seems only to flow out. But, amongst all this bull shit, there is the occasional slip of the mask, and the reality of their delicate house of cards is exposed. This happened recently with Microsoft’s CEO AI Satya Nadella, who has been pivotal in their colossal investment in OpenAI. Nadella publically stated that copyright laws needed to be changed in order to allow AI companies to train their models on copyrighted data without fear of repercussions. If this isn’t a desperate plee for help, I don’t know what is.But to explain why, we need to recap a few things first.Let’s start with why AI needs copyrighted material.All an AI does is identify trends in a dataset through a process known as “training” and then recognise these trends in other datasets or replicate those trends when prompted. Some AIs, like cancer-spoting medical AIs, only require a small dataset to become extremely accurate, as they only have one hyper-specific job to do, and their output is far simpler (positive or…