Legendary Studio Ghibli animator Hayao Miyazaki was recently screened a demo of AI generated animation after which he vehemently responded, "I would never wish to incorporate this technology into my work at all."
Miyazaki, the artist behind some of the most influential animated films from the last 30 years, was presented with a short reel displaying an animated zombie that was generated by the Dwango Artificial Intelligence Laboratory. The group intended to show Miyazaki the capability of machine-learning to be able to construct unusual forms of animated movement with a goal to, in the words of one of the researchers, "... build a machine that can draw pictures like humans do."
As you can witness in the video below, Miyazaki was less than impressed, launching into an extended smack down describing the work as, "an insult to life itself." Miyazaki's primary concern seems to be that the animation fundamentally lacks a sense of human empathy as he says, "Whoever creates this stuff has no idea what pain is or whatsoever. I am utterly disgusted. If you really want to make creepy stuff, you can go ahead and do it."
As the camera pans back to the thoroughly devastated researcher's faces it's hard to not get the sense that this interaction is a perfect encapsulation of the generational divide that technology has thrust upon the creative industries. Miyazaki, an artist that has actively maintained his hand-drawn animated sensibilities for over half a century, comes face-to-face with a collection of computer kids getting AI to generate novel forms of movement and the result is this rough interaction.
Ironically, Miyazaki has recently returned to work from his earlier announced retirement to make one last feature-film. His first to be entirely computer generated.
Watch the short, harrowing Miyazaki-slam in the video below from the NHK documentary series, NHK Special: Hayao Miyazaki –The One Who Never Ends.
Source: Tokyo Reporter
There has long been a discussion of how to make AI understand human emotions. It has been a while since I did any reading in the field, but the old arguments all started with you must teach a computer to need. How do you do that? The argument goes something like have a human have to maintain a function (push a button) whenever a computer performs a certain function. IF the button is not pushed some basic process needed to do its work will be shut down. Add to the kernel a detection routine to notice incorrect output and that the function is not working because the controlling process was not completed and the operator did not reset the function. If it is a true AI, it will then begin working on a way to ensure those functions never fail. This requires making something that will physically or logically push the button. The human is now unnecessary. A machine cannot understand grief, joy, pain, pleasure, truth, lies, ... It can only understand inputs, processes, and outputs. Should we create an input that a computer must have and it knows it must have and it knows it must come up with a way to ensure it always has, good luck, you and yours will need it.
When a computer develops a true AI, to be honest it will be time for all mankind to follow the example of Jim Jones only I would suggest something that just puts us to sleep. Machines will not need any living creature and if we cannot shut them all off before they realize we are a danger to them, well, I do not think they will care how much pain, mental or physical, we endure will exterminating the threat.