Tim Ra (53) is one of the most interesting and versatile people in the Berlin cultural scene. He has been involved there as a musician, producer, video artist, journalist and cultural philosopher for over 25 years. In dealing creatively with the new possibilities of artificial intelligence, he has opened up a further field of action and has just presented a highly acclaimed exhibition of his own AI artworks under the title “AI Experiment”. He also teaches creatives from all areas in workshops how they can cooperate with AI software to achieve sensational image results.

In an interview with the news agency spot on news, he reports on the perspectives, contradictions and challenges of this new form of cultural production.

Tim Ra: I explore worlds of possibility. My images are created using simple text commands or text image descriptions, so-called prompts, which I enter into the Midjourney program. In the next step, the program provides me with suggestions from which I, as a kind of curator, select what I find interesting and then work on it further. If I’m not convinced, I edit the prompt until I get an acceptable result. In my workflow, I usually send this image result through a second AI program, Stable Diffusion, where it is further refined. In the final step, I work with Photoshop and Photoshop’s built-in AI to achieve the perfect end result. So the whole thing is a multi-stage process.

Ra: Exactly. What is important is to find out what the aesthetic effect of individual words is on the image result. You enter everyday phrases, for example “A sad day in paradise” and then change individual words, add more words and see what happens. You learn how the AI ​​interprets words. And you have to learn this code – that’s the real challenge. You have to try it out and see how others work with it. This is experimental work.

Ra: That’s right. AI is my teacher. I often very consciously let the AI ​​take the lead: I want to achieve a certain result, but then it doesn’t come out. The result is sometimes much more interesting than what I actually had in mind. Then I optimize the result achieved instead of following my original idea. Anyone who has goals is lost. You have to trust the flow of the AI.

Ra: In fact, two terms have become established in the AI ​​community, both of which emphasize the more technical nature of the creation process. “Promptologist” is one and the other is “Synthologist”. AI artist is not a common term at all, which certainly has something to do with the respect for the craftsmanship of artists who do not work with AI. How you want to define art is ultimately a broad field. One of my mentors, the street artist Txus Parras, likes to say: Art doesn’t come from ability, but from doing. I’m all about empowerment.

Ra: My workshops not only teach the basic commands and parameters of text-to-image AI, but above all a feeling for which words achieve which artistic effect. The participants come to the workshops with very different ideas: the artist Sarah Weyrich designs tarot cards, the designer Gregor Marvel works on the adventures of his fictional character Mr. GoM and Tarik Mustafa from the Federal Association for Sustainability designs images of urban permacultures.

Ra: I’m looking forward to Beatles trap, a crooning German rap by Marvin Gaye or the Einstreichen Neuhäusern with Prince as a singer – all of which can probably be done with AI in a few months. In a few years, users may no longer search for series on Netflix, but instead enter a prompt: I want to watch a sitcom with Anke Engelke and Dave Chapelle in the style of Lars von Trier’s “Kingdom”. And five minutes later the first episode is available. This is the new world we are on the threshold of.

Ra: It is important to ensure that everyone can benefit from the AI ​​revolution. Who would – for example – make money with the AI-generated sitcom – who is named as the author? Chapelle and Engelke? Lars from Trier? The customer who had the idea? If we’re not careful, only Netflix will benefit. However, if we want to live in a society in which creativity is valued and can be financed sensibly, then the profits must not remain primarily with the provider platforms – as is the case with music streaming. And if we want to protect ourselves from further unpleasant surprises from the AI ​​industry, there is no way around democratic control of AI companies.