AI artwork often comes out surprisingly well

Artificial intelligence (AI) opens doors to new forms of art. Whether a painting is done with brush and paint or with keyboard and mouse, the essence remains the same. Art remains something that is unique. “In any case, we will have to prepare future generations for a future where AI will also play a big role in the arts.” says AI artist Scy Heidekamp.

AI systems can generate paintings as if Botticelli painted them with his own hands and compose music in the style of The Beatles. Speaking of AI-generated music, it’s been five years since the first music album produced entirely by AI was released.

Singer-songwriter Taryn Southern succeeded with the song Break Free. This song belongs to the album released in 2017 and was made with Amper software. This software can already compose, produce and much more. This way you can set a lot of preferences: the desired length, a mood, a tempo and a genre.

Cheating

AI generates works of art that humans may never be able to come up with. This is because the algorithm on which AI is based often thinks extremely ‘out-of-the-box’. That doesn’t mean people can’t. However, people’s thoughts do not always reach that far. It creates new opportunities.

However, not everyone is up for it. “There is always some fear associated with big changes,” says Heidekamp. “And it’s not surprising, because you also saw that people began to doubt the survival of their jobs when the computer was invented.”

Recently, an AI art application has received a lot of criticism. Recently, the annual art competition took place in Colorado, USA. While most participants brought a sculpture, painting or glass art, there was one exception: Jason Allen. He submitted an artificial piece of art and won the competition. It did not take long before fierce criticism arose. Artists accused him of cheating.

Appreciation of authenticity

AI makes it possible to take completely new paths of art. Not everyone is happy about this though: “What depresses me the most is how many non-artists really don’t see a problem, or even a difference, between something a human made and something an AI has ‘produced.’ If you can even talk about producing,” argues an artist on the online internet platform Reddit.

Scy understands this. She explains this with an example: “Think of an e-reader. This gadget makes it easier to download and read a book. So you might be thinking ‘why isn’t everyone doing this?’ or ‘isn’t it much more practical than reading a real book?’ But that doesn’t cover it, because enough people have continued to value ‘real’ books anyway. The same applies to art. There is always something authentic about painting a painting yourself. I think it’s something that will always be.”

Visual artist Casper de Jong, who calls himself Interactive Performance Designer, also sees more opportunities than threats in the AI ​​application. “There are certainly artists who are excited about it, but there are also plenty of skeptics. I think it is too early to judge whether AI is a threat or not. What I think is that we should better embrace and accept AI as a tool instead of positioning it as the culprit. It can do very fine things, and we must make use of that.”

Expand material package

Scy also believes that artists should better choose to embrace these tools and use them as a supplement. “Combining materials and techniques is something that will always remain interesting. I see AI as a perfect tool to expand your material package with it”, says Scy.

More and more tools are being introduced that make it easy for beginners to create art. Right now DALL-E 2 and Midjourney are the most popular. “These all have an accessible and user-friendly user interface. The technology behind it is text-to-image technology. You just have to write a text of what you want to see. Based on that, you get a final result,” explains Scy.

This tools work based on so-called diffusion models. “That is, it starts with an image that is nothing more than noise. From the noise, an algorithm is used, which has been trained by huge databases of text and photos. The diffusion model directs the noise to the description you gave it.” For example, AI algorithms create new images by combining and modifying existing images.

AI generated art of an owl – Pixabay

“Active anti-artist”

Although the AI ​​is based on other works of art, the output is unique in every case. This is due to the huge database in which millions of works of art are gathered. This is also what worries artists. As artist RJ Palmer tweeted: “This thing wants our jobs: It’s actively anti-artist.”

Scy continues: “But the story has two sides. It’s actually the data set that cares the most. For the end user, there is little information available to see exactly where this data is coming from. These datasets, on which the AI ​​learning system is based, contain huge amounts of photos and text files. The problem with this is that often artists haven’t even given permission for their work to be in that dataset.”

Fraud prevention

There is another possible use of AI in the arts: detecting fraud. It can also help detect counterfeits. AI technology can determine authenticity using non-traditional authentication methods, such as X-ray technology. The system looks at brush strokes and patterns, scans them, and can then ‘see’ in no time in which areas the forged artwork differs from the real work.

Future vision

Scy believes that the development around AI in the art industry is promising. These are still in the early stages. “In a few years, it will no doubt be different again. I think the style we use now will be completely retro in ten years. If you now look at films from twenty years ago, you see a big difference in the quality and the thinking behind it. That’s how it is in the art world’.

Leave a Comment