AI faces the event horizon

 Virtual Intelligence 

There are few topics that elicit more darkened eyes in the field of artists than that of AI; not just for what it's already capable of today, but for where we all see it heading in the very near future. We all see the rapid access the world has been granted to use quick AI models for painting and writing, but what if I told you that we're on the verge of having every major main form of art impacted by the influence of AI? What if I told you that the current stigma of "Well, it isn't the same if the human aspect of pain isn't driving the work" is already being eroded under the sheer wave of improvements that push past the perceptible human ear to counteract it. And what if I told you that there currently is literally nothing any of us can do about it in order to stem the tide of chaos?

Writing is a medium directly under attack by AI, whether we're talking about research articles or educational assignments. We're already hearing stories of students trying to forgo coursework by submitting piles of computer generated work, which fits neatly with the dry style that most non-English majors have anyway. For the time being actual creative writers are safe, and the current trajectory of AI will certainly make it difficult for models built on training to create anything in the vein of the literacy new authors possess, but when we come to bid-writing, email correspondence, technical articles, research papers; there's a very real possibility the end of those professions has already been toiled by the mere introduction of common-access models like ChatGPT. Ubisoft have already demonstrated how quickly ostensibly creative companies are to seize on the convenient.

Art is another big victim that hardly needs any introduction if you've been paying any attention to the world of StableDiffusion. Learning models built of existing work basically makes it so that any artist who does not develop a wildly unique style of their own, and perhaps covets styles that resemble certain disciplines or real life imagery- are on the route to becoming moot. Yes, for now there's still an instinctive hint to the computer generated that betrays it to the human eye, but it's a factor that is slight and being stamped out. Small efforts are being placed to try and stop this, one such is by installing a certain copyright filter that tells Learning algorithms to ignore new works; but proliferation has already overcome this. So many learning models are independently coded and simply ignore such warnings- visual art has no real means to defend itself.

Song is the other big angle we've seen of this in recent weeks. Voice training has come a huge way since it was released out into the wild, with what once to took days with thousands of audio sources being reduced to hours with only a few. The Dagoth Ur voice training algorithm is based on the performance of a character in Morrowind that has barely ten minutes of recorded dialogue, and yet people had managed to get him to read all kinds of brand new dialogues with near perfect tone and vocal variance. (With some editing on their part, obviously.) But even without creating fresh dialogue, there have been those able to mash the sound of their own voice with that of another with AI. The world was unnerved with the example of one Twitter user who created a fresh Kanye track over his own rap with nothing but a trained algorithm in order to say things that Kanye would never say- like admitting he was wrong. Imagine the effect on the music market from that alone!

Now if we were to do our best to be charitable about this, we could say that AI voices for musicians could help an artist live on after their death; but on the otherhead is that even something that we want? Some other artist wearing the voice of a person we love writing in lyrics that they think that artist would have wanted to formulate? In that Kanye example I mentioned some people already found themselves a little uncomfortable over the fact that the voice over artist wrote one line about Kanye's Deceased mother; which already clearly presents the dichotomy between what a person can say about themselves and what others can say using their mouth. Dead actors are already being brought back to retain their roles, dead artists will become a new liberty to cross for everyone.

Honestly, this whole situation brings back to mind that time when Kanye West thought it just an insanely romantic gesture to film a hologram of Kim Kardashian's dead father for her birthday. A hologram that Kanye then manipulated in order to talk about how smart her current husband was, (meaning himself) which in turn conferred her some meagre display of second-hand intelligence. Now every relationship can only be understood by the people in it, of course, but I think we can all see how objectively unhinged that is. And if it's worth anything, Kanye is no longer with Kim- so... yeah. But we're just talking about the silly applications of this replacement software, what about when we start getting to the philosophical?

Playing Cyberpunk 2077 recently has conferred to me a rather interesting parallel, in that the very idea of 'capturing the soul' of someone is a key point of the narrative. In that game, there's a special chip that is full of all the neurological data of a deceased person, allowing their loved ones to live with a virtual recreation of their long lost relative. The question rising there being whether this facsimile of the person who was counts as a real person, or is simply just a clone. I think what we're seeing of AI today isn't quite at the point yet, but it nudges along that line. What we're seeing is a near perfect mask of a person who once was, now all we need to do is to come up with a silicone husk to live behind it and soon we'll all be asking 'are friends electric'?.

So we're on the verge of a revolution of AI that threatens to redefine what it is we consider our humanity, our ability to emote with genuine intention. The entire prospect of art exists to share what it is that we feel with others around us, and if we're using computers to depict that feeling than the fabric connecting our hearts is thinning. How many layers deeper do we need to go before machines are creating the mimic of feelings that it's sharing with other machines to react to, and does that create something akin to a mechanical soul or simply a ghoulish cycle of answer-and-response that ultimately leads to nothing? And what of the free thinkers and artists of that proposed dystopia? Grim tidings await our lot, mark my words.