First wave of AI Art

At the moment we are in the first wave of philosophical debate of what the legal ramifications of using AI to create art. (and yes, I use AI for my art creation – it speeds it up enormously). What is getting lost in the debate, maybe, is that the AI behind this isn’t recreating the artists’ works. It is processing input to perfect its process. The AI is “trained” using artists’ works – hopefully with the permission of those artists or through using legally allowable stock imagery – so that it can interpret other input (what is being given to it) in a meaningful way. Much like any of us do when we sit at a gallery and copy an artwork on the wall to get a sense of how the artist created the work in front of us. We aren’t stealing or plagiarizing the image, we are looking at the process of making that image. We may duplicate that artwork during the search for the process – but we don’t claim it as our own (that would be forgery). Yet things we create after that initial action may use the creative process we discover through recreating the image before us in the gallery – you might paint ‘in the style’ of Van Gogh, for instance, and we can be considered to have been trained in that way, but any artist moves on. What ‘new processes’ we learn we adapt to our way of doing things to our artistic goals. The AI is like this, seeking a process, not an image. The AI is deciphering and perfecting (the original is not destroyed, nor “appropriated” – it is “discussed” by the AI algorithm). The AI’s algorithm, ever-changing, uses the interpretations it gets from all its past experiences to present alternatives to the input we give it, and thereby ‘create’ a new “piece of work” that at the time of creation has no ‘value’. We take the options and give value to it by using it.

Here’s the other thing – The AI doesn’t make the input. It doesn’t ‘create’. It’s like those old time secretaries who takes our scribbled notes and types it up for us on those old clattering typewriters. We give it the ‘source’ and it gives us the tidied-up bits. We guide it to do what it does, and it does it based on its learnt experiences (its magical “algorithm”). For us, the creators who use it, that isn’t the end of the process either. The AI doesn’t YET get it right the first time – which means we tidy it up and retrain the AI, and the AI OUTPUT is only the first step in the process, and it’s only one small piece in the whole workflow to create something much larger than the parts, which comes from human thought and human decision making and human sensibilities.

As I said, we’re in the first wave of the philosophical processes, and there are going to be many voices raised for both sides of the debate. BUT I’m not waiting around for it to all be sorted out – illustration and film making (including animation) is a dang long-winded process and anything that helps is worth working with. But always Create! Tell your stories the way you want to tell them, using what can get closest to that inner vision you have.