
First of all a disclaimer: my expertise in art(s) is as close to “zero” as it can possibly be. I can tell if I like something but I am not able to tell the difference between a Kandinsky and my young grandchild scratch & scribble. In other words, I am not to be trusted if I speak of art (and I I usually never do…).
Having said that, I am impressed by the AI art generator in its ability to create a “piece of art” starting from two hints: a sentence (even a single word would do) and a style. The best you can do to see what I mean is to try it out. In addition to the link you can also download the app for your tablet and smartphone (Dream by Wombo, available for free on both Android and IOS store). It is both fun and, at lest to me, impressive.
I played a bit with it and noticed that every time it comes up with a different “rendering” even if you start with the same input. However something remains unchanged, like, “love” gets associated with a couple, rendered in different ways every time (and yes, it is always a she and a he, AI, at least this one, is not gender neutral).
AI and arts are no strangers. I have read of very many application of AI in the art field. One that impressed me is the use of AI to restore the Rembrandt’s masterpiece, the Night Watch, at the Amsterdam museum. It was painted in 1642 but when it was hung in the Amsterdam Townhall in 1715 the people in charge discovered that it was too large to fit on the wall and so they decided that the easiest thing to do was to trim it to make it fit. The parts that were cut were never retrieved but the Amsterdam Museum has resorted to AI to recreate them. You can read the story and the details of the use of AI for reconstructing the original here. It makes for an interesting reading.
Hello Roberto,
indeed, fascinating to play with. For example, entering “Ferrari”, you get a surreal painting of a Ferrari, even with horses in the background. Is it art or not? Good discussion. As art should have a message, maybe it is not; at least not for the average user. Another interesting question, as the tool must have learned from picture database, what about copyrights.