You might have seen my recent blog post titled “Raspberry Beret(s),” which is an article that recommends some of the best songs to queue up in your headphones on your next thrifting haul. If you haven’t read it yet, which I would definitely recommend giving it a leaf through, it features a collage of remixed portraits of women wearing raspberry berets (yes, like the song by Prince) in the iconic pop art style of Andy Warhol. Now, I definitely did not draw this detailed group of figures by hand, instead I used a relatively new technology known as an artificial intelligence (AI) image generator.
To create these images I used Picsart’s free text to image AI generator, which is extremely similar to the more well-known DALL-E system. I had heard a lot of buzz going around regarding this new-fangled AI image generator programs – how cool they are, how easy they are to use, how they might be infringing on copyright or people’s privacy, etc… – so when the chance to test drive one of these babies was thrown in my face I knew I just had to take it.
Picsart, which is usually my go-to free alternative to Adobe Photoshop, is really quite intuitive and pretty straightforward with all of its features, and there was no exception for its AI generator. To get an image, all I had to do was type in a prompt or phrase into the search bar and in a click of a button, its AI engine took care of the rest, however, much like DALL-E, the images’ “success rate can depend on how the caption is phrased (https://openai.com/blog/dall-e/). It did request for me to provide a moderate amount of detail in my prompt, and it was also recommended that I select a handful of key words from a cache of various adjectives. I suppose the more information you give to the AI, the more precise and personalized it can get with its images, which I was a bit surprised by because I had assumed that too much information might overwhelm its systems and result in a sloppy mashup of digital goop. I settled on the prompt “woman wearing a raspberry beret in pop art style” along with the key words: Andy Warhol, illustration, portrait, and vibrant colours. The results were really quite impressive – if you had shown me those images and told me that Andy Warhol created an entire raspberry beret series of pop art works I would have absolutely believed you.
So, since I went to all this difficult and tiresome work to type this all out to create the images, do you think that it is fair to consider myself the artist of these pop art portraits? I honestly don’t think so. At the end of the day, I think that images generated by AI systems such as DALL-E and Picsart aren’t really art – the AI image generator is more of a tool to help us humans conceptualize and visualize different and perhaps abstract ideas. The AI makes images out of other images, rather than using images to pull reference from or inspire new images or works like organic artists do. So, please don’t start thinking I’m some visionary digital artist with a whole bunch of Andy Warhol flare – I’m just capable of pressing some keys on my laptop to form a sentence to get some magical robot to draw me things.
- Antonelli. William. 2022. How to use DALL·E mini, the viral AI tool that can turn any prompt into a series of pictures.
- 2021. DALL·E: Creating Images from Text