I haven’t produced any tangible artefacts, documents, or other evidence of being productive this week, but that doesn’t mean I haven’t been thinking. In fact, I’ve been thinking a lot.
This week, I was pretty similar to The Thinker, sat stationery and pondering.
It’s easy to feel unproductive when your work, writing and research are not in ‘real’ motion. But moments of stillness and reflection are equally important. Some of my best revelations come when I close the laptop, put the books down and sit with the ideas, theories, and possibilities that live inside my mind.
I’ve been particularly consumed with the concept of generative AI, a type of artificial intelligence that can generate new content such as images, text, or audio based on a set of inputs or a learned model. This week, I set about exploring ways to effectively use these tools in my everyday workflows and push their limits (for research purposes, of course). And I observed something very interesting. When it comes to generative AI tools, you get out what you put in.
By this, I mean that Midjourney and ChatGPT sometimes met basic prompts with unoriginal or uninspiring results. Other times, the AI’s output differed from what I envisaged when I entered the prompt. For example, Midjourney did quite well with brief prompts (like the featured image of this post – ‘cinematic still shot, a blog post featured image for an article about the future of AI‘ –v 4 –ar 3:2‘), and photo mashing (top row), but seemed to struggle with reproducing my face (bottom row).
The process of trial and error to achieve my desired output was time-consuming and sometimes frustrating, especially given the promises of generative AI tools that I had seen in news articles, tweets, and TikToks. It seemed there was an art to crafting the perfect prompt as each AI system learned, conversed and created differently. This discovery reminded me of a statement from Lev Manovich’s ‘Defining AI Arts: Three Proposals‘:
There are at least three points in this process where a human author makes explicit choices and controls what computer would do. First, a human designs network architecture and also an algorithm used to train a network (or selects from the existing ones). Second, the human creates the training set. Third, the human selects what in her/his views are most successful artifacts from many more the network generates.
I agree. While these tools are extremely powerful, it does come down to a human operator’s creativity, skill and knowledge at the end of the day. So, I started thinking about AI as an artistic medium rather than a tool or software. As something that we use to create art rather than something that is autonomously generating art for us.
And, because I am a perfectionist, I was determined to learn how to use Midjourney effectively. Soon I was down the rabbit hole of Discord prompt chats, informative Twitter threads, and entire databases dedicated to teaching users what to tell Midjourney to generate an image of a particular style or feature a specific subject. In this manner, navigating the learning curve of generative AI entails engaging with a participatory culture.
A participatory culture is a culture in which members actively participate in creating and shaping the culture they are a part of rather than simply consuming it passively. Its critical identifier is the social relations that emerge when audience members collaborate while appropriating and remixing media to produce new meanings. In this way, learning how to use the medium of generative AI is a participatory culture in which users actively participate in remixing media to create new texts. This is evident in the communities formed around the use of Midjourney, where users engage in frequent knowledge-sharing and collaboration to produce digital art. My Twitter feed is flooded with threads instructing Midjourney users on what prompts to use to achieve specific outputs and sharing new techniques for AI image generation (which I am very thankful for).
I find this fascinating. Because the artistic medium of AI is so new, there are no art schools that ‘teach’ it, no bachelor’s degrees to be obtained in Generative AI Design, no Midjourney user manual, and no AI art history (yet). We are all learning from scratch and learning from each other in an unprecedented and unique participatory culture.
As Lev Manovich (2018) believes, AI plays a crucial role in our global cultural ecosystems, and its function will only get more expansive. Are you going to ignore it or embrace it?
All images in this article were generated using Midjourney.
ChatGPT kindly assisted with the editing process of this blog.