All over the web, we are witnessing very spectacular results from statistic algorithms that have been in the work for the last forty years. We gave those algorithms an incredibly catchy name: "Artificial Intelligence". [...] All of this has been made possible because billions of humans were uploading and sharing texts and pictures on the commons we call "the Internet" [...] What we are witnessing is thus not "artificial creativity" but a simple "statistical mean of everything uploaded by humans on the internet which fits certain criteria". It looks nice. It looks fantastic. While they are exciting because they are new, those creations are basically random statistical noise tailored to be liked. [...] But one thing is happening really fast. Those "artificial" creations are also uploaded on the Internet. Those artificial artefacts are now part of the statistical data. [...] The algorithms are already feeding themselves on their own data. And, as any graduate student will tell you, training on your own results is usually a bad idea. You end sooner or later with pure overfitted inbred garbage. Eating your own shit is never healthy in the long run.
And consider the following comment from Niti:
Also consider the following comment from Zach: