cover ilustration prompt> image-making is an ai prompt by Adobe Firefly
When image-making becomes a prompt,
what are we really teaching?By Ivy Wei, MA Researcher, Goldsmiths, University of London
Over the past few years, generative AI tools like Midjourney, DALL·E
have rapidly entered the creative classroom. From primary school
workshops to university design modules, students are now creating
visual work by typing prompts into algorithms. A few seconds later, a
polished image appears.
This signals a democratic shift in creative access for some. For others, especially educators in traditional visual arts disciplines, it raises uncomfortable questions: Is this still artistic training? What are students actually learning? And more importantly—are educators prepared for this shift?
From hand and eye to prompt and click
Traditional visual arts teaching has always valued physical engagement: observing from life, exploring materials, drawing by hand. This process is slow, sensory, embodied, and deeply tied to a student’s lived experience. Through repetition, failure, and improvisation, students learn not only to create, but to feel through making.
In contrast, AI-based image generation bypasses the hand and eye. It is fast, seemingly effortless, and heavily stylised. A student can type “a glowing jellyfish floating through a coral forest” and receive different sophisticated results in seconds. But in doing so, what is being exercised? Imagination, or search logic? Interpretation, or satisfaction?
Some artists and educators have noted that integrating AI tools into visual arts education is not just about learning new software, it fundamentally challenges the assumptions of what “making art” means. As one contributor put it, many art teachers trained in traditional methods feel less like they are extending creative tools, and more like they are witnessing a structural disruption in the classroom.
Are we still teaching “how to make”, or just “how to choose”?
In a generative AI workflow, the artist’s role changes. Instead of constructing an image step-by-step, they now describe an idea, receive visual options, and decide what to refine. This process more closely resembles editing than drawing.
Some artists and educators describe this shift as moving from “maker” to “curator of possibility.” The artist no longer owns the marks, they own the framework, the intention, the iteration. Yet this dual role—both as prompt designer and image selector—requires an entirely new set of skills. Students must now learn how to formulate powerful prompts, read visual results critically, recognise bias and repetition, and refine with purpose.
This demands a different pedagogical mindset. Teachers are no longer only demonstrating techniques, they are teaching students to navigate systems, frame decisions, and define value in a context where the machine generates at overwhelming speed. As one educator from the project noted, “AI doesn’t know what’s good. It only knows what’s likely.”
Teachers as curators, not just demonstrators
Some projects have begun to explore how visual arts teachers can engage with AI in meaningful, critical ways. The aim is not to master every tool on the market, but to help students understand how AI works, what assumptions it carries, and where human agency still matters.
Some proposed strategies include:
- Comparing AI-generated imagery with hand-made work to analyse compositional depth and nuance
- Asking students to “remix” or “correct” AI outputs as a form of active authorship
- Tracing the visual stereotypes embedded in generative datasets: who is represented? and who is not?
- Encouraging students to build their own mini datasets or intervene in the system’s logic to make their values visible
Guiding students through hybrid making processes that combine imagination, prompting, and hands-on skills: in one example, children were encouraged to design a robot with the help of AI. The teacher asked them to brainstorm prompts describing why such a robot was needed, then sketch their ideas by hand, and finally use AI tools to generate a 3D model. (Figure 1, 2)
This turns the classroom from a site of software use into a space for critical visual inquiry. Teachers become cultural translators and systems thinkers, helping students understand not just how to use tools, but how to position themselves in relation to them.
The myth of creativity and the politics of tools
AI’s rise also forces us to revisit long-held assumptions about creativity itself. What does it mean to create something original? If a student uses a tool trained on thousands of past artworks, how do we distinguish between homage, replication, and invention?
Visual artist Trevor Burgess has argued that many AI systems resemble collage machines rather than true creators. “It seems closer to collage than anything generative… what the AI seems to do is stick things together in a very sophisticated way,” he notes. Unless an artist codes their own system or deeply manipulates it, they are ultimately working with someone else’s aesthetic parameters.
This raises uncomfortable but necessary questions: Who designs these systems? Whose ideas of beauty, realism, or coherence are embedded in them? If students are not made aware of this context, they risk internalising styles and structures without knowing where they come from.
More access, more dependence?
There is no doubt that AI makes image-making more accessible. For beginners, children, or individuals with physical limitations, it offers an exciting entry point. Many students who previously felt alienated by technical skill barriers can now express ideas more freely.
However, this accessibility also brings a new kind of dependence. Without visual literacy and critical framing, students may come to rely on default aesthetics, algorithmic tropes, and instant visual gratification. They may mistake fast results for meaningful ones. They may stop asking what a picture means, and only ask if it looks “good enough.”
One risk is that the process of making art shrinks into a process of selecting and fine-tuning what the machine offers. The core of art, which is ambiguity, physicality, and friction, risks being flattened into an infinite scroll of stylised outcomes.
Rebuilding educational foundations
The challenge is not just technical, it is institutional. Many art schools, especially at the secondary and foundation level, are built around craft-based curricula: drawing, printmaking, ceramics, observational painting. AI doesn’t replace these directly, but it reframes their position.
Should students learn drawing before using AI? Or after? Should AI tools be part of the foundational curriculum, or introduced in critical theory seminars? Should students be assessed on the quality of their AI outputs? Or on how they use the tool to support artistic vision?
These questions are pedagogical, philosophical, and ethical. They require institutions to re-examine not only their content, but their values.
In an age of infinite visual generation, the most important lesson may no longer be how to draw, but how to decide. Not what to make, but why to make it.
Art education must help students become not just users of tools, but reflective shapers of culture. In this new landscape, art teachers are not simply instructors. They are translators of visual systems, facilitators of critique, and guides in a changing world of creativity.
So, is visual arts education ready for AI?
Not fully, but the conversation has started.
