Ivy Wei, MA Arts Administration and Cultural Policy (Goldsmiths)
Two Ways of Seeing
Imagine two scenes. A century ago, Monet spent months watching light move across the haystacks in a field. He observed them through fog, dawn and sunset, seasonal change. The act of painting took time, but the deeper work happened before he touched the canvas. In 2025, a designer types a short prompt into an AI generator. In thirty seconds she receives numerous images of haystacks in every season. She generates variations, then more variations. After several hours she feels drained and unsure what to pick.
Speed gave her more images than Monet ever painted, yet somehow she feels slower. This is the speed trap. AI promises instant production, but many creators report that their workflow has become more tiring and less focused. The question is simple: has speed made us more efficient, or has it rearranged creativity in a way that works against us?
How Speed Changes Creative Work
1. Time Is Redistributed: From Making to Choosing
AI tools have not removed creative labour, they have shifted where it happens. Traditional art placed time in the making (execution) while AI places time in the choosing (selection and evaluation). The human hand has been replaced by the human eye. Monet’s long observation helped him understand seasonal patterns. The AI user receives all seasons instantly, but without the gradual process of looking that once shaped perception. The images exist, the embodied understanding does not. This is known as the lack of “embedded understanding” – meaning the deep cognition that is accumulated through bodily practice and experience.
When the number of images increases, the ability to judge them decreases. Your brain starts asking the same questions again and again. Is this the best one? Should I try one more variation? Would a new prompt give me something even better? The cycle becomes addictive and tiring.
2. Decision fatigue becomes the new bottleneck.
Choosing between options consumes finite cognitive resources. When options increase, fatigue increases too. This phenomenon is known in psychology as “ego depletion”, proposed by Roy Baumeister et al. It suggests that willpower or decision-making ability depends on limited resources. Small studies on AI-assisted ideation suggest a related concern. Research in the field of Human-Computer Interaction (HCI) has shown that relying too early on image generators in the early concept design phase can distract attention resources, shifting them to cue word management and result filtering (i.e., constant comparison and prompt management). Some experimental groups consequently produced less diverse concepts because they were trapped in a cycle of evaluation and adjustment.
The designer scrolling through hundreds of haystack images is not experiencing abundance. She is experiencing judgment overload, a state where the mind is locked in constant evaluation. Creativity requires periods of unfocused thought, yet instant generation keeps the mind locked in constant evaluation.
Three Strategies for Reclaiming Control
1. Strategy for Focus: Build Delay and Space into the Process
Paradox: Fast production can interfere with slow thinking.Creativity doesn’t solely depend on rapid output. Neuroscience suggests that insights often arise when the brain is not focused on a task. This “resting-state activity”, a relaxed or wandering state of the brain, is directly related to memory consolidation and internal exploration. If we don’t allow pauses between “idea” and “output,” we reduce the opportunities for these deep insights to form.
How to apply it:
24-Hour Rule: Generate images today, decide tomorrow. Many creators find that their preferences and judgments of quality change after a night’s sleep.
Batch Generation: Generate a fixed number of variations in a single session and review them several days later. Separating production and evaluation effectively reduces the mental stress of doing both simultaneously.
Educational Applications: Require students to complete concept notes or hand-drawn sketches before engaging in AI generation. This ensures that their thinking is fully developed before the tools accelerate execution.
2. Strategy for Depth: Combine Fast Moments with Slow Phases
Paradox: Instant feedback weakens long-term value.Social media rewards speed, rapid production leads to rapid release. However, rapid visibility is not equivalent to lasting value. When creators work primarily to meet immediate metrics, they may substitute thoughtful development with rapid content cycles.
How to apply it:
Slow First, Fast Later: Employ an alternating workflow. For example, artist Anna Ridler first manually photographed and labeled thousands of tulip images (slow phase), then trained a model to generate new images (fast phase). The final work bears traces of both.
Protect Intent: Allow AI to accelerate variations and execution, but the slower phases of intent and interpretation must be protected.
3. Strategy for Meaning: Treat Time as Part of the Artwork
Paradox: Time is removed, meaning is diluted.The value of generative art often lies in the “slow design” behind it. For example, Tyler Hobbs spent months developing the algorithmic space for Fidenza. While the output is instantaneous, its meaning is built upon extended development.
Hoe to apply it:
Making Duration Visible: Recording timestamps, iteration counts, or hints of history as part of the work.
Long-Term Projects: Creating projects that evolve daily or monthly. In this case, duration is not eliminated by AI, but rather becomes visible and important in another form.
Conclusion: Rethinking What Efficiency Means
AI has made it faster to create images. It has not made it faster to reach insight. For many creators, the quickest path to meaningful work involves choosing when to pause. Meaningful slowness is not a rejection of technology. It is a way to reclaim control over pace. Speed is useful when entering new possibilities. Pause is useful when deciding what matters.
Monet waited for seasons to change. We no longer need to. But the principle behind his patience still matters. Not because slowness is virtuous, but because thought needs room to unfold. If you want to test this for yourself, try a small experiment. Generate a batch of images today. Put them aside. Return tomorrow and choose again. Notice whether your sense of quality shifts after time has passed.
The exit from the speed trap is not to slow the machine. It is to slow ourselves.
Reference
Barbican Centre, 2019. Anna Ridler: Can datasets create art? Google Arts & Culture. Available at: https://artsandculture.google.com/story/anna-ridler-can-datasets-create-art-barbican-centre/_gXholnI1pkrLg?hl=en [Accessed 24 Nov 2025].
Chen, L., Song, Y., Guo, J., Sun, L., Childs, P. & Yin, Y., 2025. How generative AI supports human in conceptual design. Design Science. Available at: https://www.cambridge.org/core/journals/design-science/article/how-generative-ai-supports-human-in-conceptual-design/6B60B59DF0CBE764A94E484DD835F73F [Accessed 21 Nov 2025].
Chuang, Y-T., Chiang, H-L. & Lin, A-P., 2025. Insights from the Job Demands-Resources Model: AI’s dual impact on employees’ work and life well-being. International Journal of Information Management, 83, 102887. Available at: https://www.sciencedirect.com/science/article/pii/S0268401225000192 [Accessed 25 Nov 2025].
Hobbs, T., 2021. Fidenza. [online] Available at: https://www.tylerxhobbs.com/words/fidenza [Accessed 25 Nov 2025].
Doshi, A. R. & Hauser, O. P., 2024. Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, vol.10, issue 28. DOI:10.1126/sciadv.adn5290. Available at: https://www.science.org/doi/10.1126/sciadv.adn5290 [Accessed 25 Nov 2025].
Kim, B-J. & Lee, J., 2024. The mental health implications of artificial intelligence adoption: the crucial role of self-efficacy. Humanities and Social Sciences Communications, 11, article no.1561. Available at: https://www.nature.com/articles/s41599-024-04018-w [Accessed 25 Nov 2025].
Muehlhaus, M. & Steimle, J., 2024. Interaction Design with Generative AI: An Empirical Study of Emerging Strategies Across the Four Phases of Design. arXiv. Available at: https://arxiv.org/html/2411.02662v1 [Accessed 23 Nov 2025].
Oppenlaender, J., Linder, R. & Silvennoinen, J., 2024. Prompting AI Art: An Investigation into the Creative Skill of Prompt Engineering. International Journal of Human-Computer Interaction, 41(16), pp.10207-10229. Available at: https://www.tandfonline.com/doi/full/10.1080/10447318.2024.2431761 [Accessed 23 Nov 2025].
