AI and art
I am wrestling with the ideas of how one might train an AI without the risk of plagiarism, and my general sense is that when people use an AI explicitly to create something "in the style of" something or someone else the risk of the AI using unaltered training data as a result is higher. When you push an AI to try to use it skills on unexpected material (as I try to) the experimentation and frustration of the plagiaristic impulses hopefully begins. The critique of the images and the tools can also begin. I am not at all offended or put off by the criticisms of my work based in the ethics (perceived or otherwise) of the app I am using, but I feel like the work I am making with it is designed to open up an expansive discussion of how any intelligence (artificial or otherwise) learns to make art without some level of copying.
I will never defend how Lensa or any other corporate monolith runs their business, but I do think there is a segment of the "creative" community who are so concerned with notions of "copyright" that they can be stifling to genuine experimentation and creativity. The more I read about Lensa, the less I am interested in using their product because of the ethics of the company. I do not think that discounts the idea that using one’s own original work to filter through an AI is inherently theft or devoid of creativity regardless of how the AI is trained. I would never intentionally present images created by an AI that was a direct copy of the work of anyone else.I also would ask, at what point does a Photoshop filter’s effects on an image function differently than an AI’s manipulations. This is not a flippant question. I think the better we answer that question the more we are able to hold digital tool makers accountable for how they develop their tools. The process of editing images collected with an aesthetic or conceptual lens can be just as creative as drawing, painting or sculpting using a chosen subject or aesthetic or conceptual approach. There are very talented renderers with styles more derivative than any AI I have seen, and there are some more conceptual artists who have pushed the ideas of AI in more productive directions through their unique forms of experimentation. The reverse is also true. I am interested in showcasing the thermal camera I built, and the images I have made from it. Because thermal images are basically a data set that can be rendered with any assigned pallet, it seemed natural and interesting to treat these images as a data source for an AI. It also was for me a response to the fact that so many people were posting selfies filtered by Lensa all over social media. The moment may be past now, and i will look to more ethical and interesting tools. Folks may find the images I have presented ugly, or unappealing for a variety of reasons. That is fine. Everything I have presented here has been based in my own unique source material. Any “theft” done by Lensa to create their app, I was unaware of at the time. I am also still researching. I thank you all for drawing my attention to a broader understanding of the problems of using AI, though the vitriol seemed a bit unwarranted. I am sure I will continue to get some, but hope that some of you will understand I mean no harm and am trying to develop my work in a manner that is experimental, conceptual, aesthetically challenging, and ethical. I am sure of my ethics, not so sure of Lensa (or frankly any corporation)′s ethics. I may be using the “Blaze” button on this post in hopes of reaching my critics, but will be less prone to use the “Blaze” function for my work in the future.