This year, we collaborated with Kingston School of Art to give MA students the task of creating their own better images of AI as part of their final project.
In this mini-series of blog posts called âBehind the Imagesâ, our Stewards are speaking to some of the students that participated in the module to understand the meaning of their images, as well as the motivations and challenges that they faced when creating their own better images of AI.
In our second post, we go âBehind the Imageâ with Minyue about her piece, âLet AI Become Your Magic Wandâ. Minyue wants to draw attention to the overlooked human input in AI generated art and challenges those who believe AI will replace artists.
‘Let AI Become Your Magic Wand’ is not available in our library as it did not match all the criteria due to challenges which we explore below. However, we greatly appreciate Minyue letting us publish her images and talking to us. We are hopeful that her work and our conversation will serve as further inspiration for other artists and academics who are exploring representations of AI.
Can you tell us a bit about your background, and what drew you to the Better Images of AI project at Kingston School of Art?Â
Minyue is from China and previously studied a foundation course in the UK before starting the Masters in Illustration at Kingston University. Before starting the Masters, Minyue had limited knowledge of AI and she only saw discussions about it on social media – especially from artists fearful that AI tools were capable of copying their own work without their consent. At the same time, Minyue also saw many fellow creators developing impressive works using AI generator tools – whether in the ideation phase or to create the final artwork.
Confused about her own perception of AI, Minyue was drawn to the Better Images of AI project to learn more about the relationship between humans and AI in the creative process.
Could you talk us through the different parts of your image and the meaning behind it?
Minyueâs piece is focussed on two halves of a star. One half is called the âevaluation half starâ which represents AIâs image recognition capabilities (the technical term is the âDiscriminatorâ). For Minyue, recognition capabilities refer to AIâs ability to interpret and understand input data. For image generator tools, AI systems are trained on vast amounts of imagery so that they can identify key features and elements of a picture. This could involve recognising objects, styles, colours or other visual aspects. Therefore, in generating an image of a chick (as shown in Minyueâs image), the evaluation half star is focussed on interpreting what distinctive features the training data classifies as a true representation of a chick – like perhaps the yellow colour and the shape of a beak.
The other half is called the âcreation half starâ which portrays the image construction capabilities of AI tools (the technical term is the âGeneratorâ). The Generator enables AI to create new, coherent images based on the evaluation half starâs understanding of input data.
Therefore, together, Minyueâs image shows how the half stars make a full star – capable of generating AI art based on user prompts and trained by vast image datasets. Youâll see that in the bottom part of Minyueâs image in the computer tab, she indicates that the full star (consisting of the creation and evaluation half stars) make up a magic wand when added with a pencil. The pencil symbolises the human labour behind the training of both the evaluation and creative half stars.
Without being guided by humans, Minyue believes that these two half stars would not exist. It is humans that have created the input data, it is humans that prompt AI tools to create certain images, and it is humans that train the AI systems to be able to create these images in different ways. Therefore, her piece highlights the crucial human element of AI art which is often overlooked.
Lastly, Minyue also hopes to emphasise that the combination of these AI tools with humans offers a new avenue for realising human creativity. That is why she has chosen to use a wizard and magic wand to depict how AI and humans, when working together, can be magical.
Better Images of AI aims to counteract common stereotypes and misconceptions about AI. How did you incorporate this goal into your artwork?
Minyue emphasised that the main misconception that she wanted to focus on is that AI is a tool requiring rational human use, rather than an autonomous creator. When looking at her work, Minyue wanted people to contemplate, âwho is controlling the magic?â, and prompt us to think more carefully about the role of humans in AI art.
What was your motivation/inspiration for creating âLet AI Become Your Magic Wandâ?
Firstly, as an illustration student, Minyue was particularly interested in the role of AI in the creative industry. The metaphor of the magic wand comes from her observation of artists who skilfully use new technologies to create their work, which made her feel as if she were watching a magical performance.
Secondly, Minyue wanted to raise awareness to the fact that using AI image generators still requires human skill, creativity and imagination. A wizard can only perform magic if they are trained to use the wand. In the same way, AI can assist artists to create, but artists must learn how to use this technology to develop innovative, appealing, and meaningful works of art.
Finally, she hopes to dispel the idea that AI art will limit creativity or the work of human artists – instead, if creators choose to work with AI, it could also enhance their capabilities and usher in a new genre of art.
Based on Better Images of AI criteria for selecting images, we had to make the difficult decision to not upload this image to our library. We made this choice based on closer scrutiny of the magic wand metaphor which could be misconceived as promoting the idea that AI is magic (this rhetoric is commonly pursued by technology companies).
What do you think of this feedback, and was this idea something that you ever considered in your process?
Minyue understood the concerns and appreciated the feedback provided by Better Images of AI which made her reconsider how her work could be misleading in some aspects and the challenges of relying on metaphors to communicate difficult ideas. Her intention was that the magic wand metaphor would prompt individuals to think more deeply about who is in control of AI art and also, how AI can advance the creative industry if used safely and ethically. However, she is aware that coupled with the technology industryâs widespread use of magical symbols to represent AI (for example, the logo for Zoom’s AI Smart Assistant or Google’s AI chatbot Gemini), Minyueâs image could (unintentionally) be perceived to suggest that AI (alone) is magical.
Was there a specific reason that you focussed on dispelling misconceptions about the human element of AI art, especially in relation to image generation?
Minyue strongly believes that the creative power of AI comes from human inspiration and human creativity. She hopes her work will help convey the fact that AI art is rooted in human creativity and labour, which is often overlooked in the public discourse promoted by the media about AI replacing artists, leading to misunderstandings.
A lot of the inspiration for Minyue holding this view has come from her reflections on how past technology has integrated into the creative industry. For example, painters were originally fearful about the widespread adoption of photography since it offered a faster and cheaper means of reproducing and disseminating images. But over time, Minyue believes that we can see how photography has developed its own unique styles and languages, with photographers moving away from imitating traditional art pieces, to explore unique photographic expressions. Minyue believes that AI may also evolve into a new tool for the production of a new art form.
Can you describe your process for creating âLet AI Become Your Magic Wandâ?
Minyue detailed the very long process that led her to the final creation. She recalled how having the Better Images of AI Guide was helpful, but she still struggled because her initial understanding of AI was really poor.
Therefore, Minyue took time to carefully research the more technical aspects of AI image generation so she could more accurately represent how AI image generators work and their relationship with human creators. Below you can see how she researched the technical elements of AI image generation as well as its use in different contexts.
Minyueâs initial sketches also show how she was interested in portraying the relationship between humans and technology.
Minyue aims to create more engaging and approachable AI images to help non-experts understand AI technology and reduce public fear of new technologies. This was also one of her reasons for choosing to participate in the Better Images of AI project.
What have been your biggest challenges in creating a better image of AI? Did you encounter any challenges in trying to represent AI in a more nuanced and realistic way?
Minyue faced difficulties when challenging her previous views on AI that were presented to her by the media. Contrary to a lot of the other images in the Better Images of AI library, Minyue also wanted to promote a more optimistic narrative about AI â that AI can be beneficial to humans and enhance our own creative outputs.
Another one of the challenges that Minyue faced was distinguishing between AI and computers or robots. In one of her initial sketches she shows how in her early stages of this project, Minyue overlooked how AI has numerous applications beyond just being used within computer applications.
What have you learned through this process that you would like to share with other artists or the public?
Minyue says that while artists are often driven by their passions when creating their works, it is important to consider how art might cause misunderstandings if creators are not guided by in-depth research and detailed expression. Minyueâs hope is that other artists will focus on this in order to promote a more realistic and accurate understanding of AI.
Minyue Hu (she/her) is about to graduate from Kingston University with a Masterâs degree in Illustration. In the coming year, she will be staying in the UK to continue her work as an artist and actively create new pieces. Minyue’s inspiration often centres on human experience and emotion, with the aim of combining personal stories with social contexts to prompt viewers to reflect on their own experiences. Her final project, Daughters of the Universe, is set to be released soon, and she looks forward to your attention.