Designing with the Machine: Rethinking Creative Agency

by

Author: Denis Avetisyan

As artificial intelligence becomes increasingly integrated into the design process, understanding the shifting dynamics of creative control is paramount.

This review examines the interplay of agency between designers and emerging intelligent technologies, highlighting the importance of designer introspection and adaptive expectations for maintaining creative empowerment.

Despite the promise of enhanced creativity through collaboration, designers working with increasingly powerful intelligent technologies often find their agency subtly undermined. This paper, ‘An Intent of Collaboration: On Agencies between Designers and Emerging (Intelligent) Technologies’, investigates the dynamic of creative power between designers and large language models (LLMs) through an experimental co-creation process. Our findings reveal that maintaining creative empowerment requires designers to cultivate self-awareness of their process, critically understand the capabilities of the technology, and deliberately adjust the working relationship. How can we foster collaborative ecosystems where designers and intelligent technologies mutually amplify, rather than diminish, creative agency?

Reclaiming Agency in the Act of Making

Contemporary design methodologies frequently concentrate on the tangible result, the finished product, to the detriment of recognizing the profound learning embedded within the iterative act of making itself. This outcome-oriented approach often overlooks the subtle insights, material understandings, and unexpected discoveries that emerge as designers grapple with form, function, and fabrication. Consequently, valuable knowledge-tacit understanding of materials, embodied skills, and nuanced problem-solving strategies-is frequently discarded or undervalued, as the emphasis remains fixed on delivering a polished endpoint rather than cultivating a deeper, more holistic comprehension of the design process. This prioritization of outputs can stifle genuine innovation, limiting the potential for designers to leverage the full breadth of knowledge gained through hands-on exploration and iterative refinement.

The prevailing focus on finalized design outputs frequently overshadows the critical importance of designer agency and the development of embodied understanding. When evaluation centers solely on the finished product, the iterative process of making-the hands-on exploration, the intuitive adjustments, and the tacit knowledge gained through material engagement-is devalued. This prioritization not only limits opportunities for genuine innovation, as designers are incentivized to quickly reach a predetermined outcome, but also hinders the cultivation of a deeper, more nuanced understanding of design problems. Consequently, the ability to respond effectively to unforeseen challenges or to leverage unexpected material properties is diminished, ultimately impacting the richness and adaptability of the final design.

The pursuit of innovation isn’t solely about achieving a desired outcome; it fundamentally relies on the knowledge cultivated through the act of making itself. Approaches like ‘Research Through Design’ recognize this, prioritizing iterative exploration and embodied understanding over a linear progression towards a finished product. This methodology frames the design process not as a problem to be solved, but as a means of inquiry – a systematic investigation where failures and unexpected discoveries are valuable data points. By valuing the ‘how’ alongside the ‘what’, designers can unlock deeper insights into materials, techniques, and user needs, fostering genuinely novel solutions that extend beyond incremental improvements and truly redefine possibilities. This shift encourages a more responsive and adaptive practice, allowing innovation to emerge from a continuous cycle of learning and refinement.

Navigating the Shifting Power Dynamic with AI

The integration of artificial intelligence tools, such as Midjourney and Canva Image Generator, into textile making processes is fundamentally altering the established relationship between designers and technology. Traditionally, designers exerted primary control over creation, utilizing tools as extensions of their direct input. However, these AI platforms operate with a degree of autonomy, generating outputs based on prompts and algorithms, which necessitates a shift in how designers approach creation. This isn’t simply a matter of designers learning to use new software; it represents a redistribution of creative influence, requiring designers to manage and refine AI-generated content rather than solely producing work from inception. Consequently, the power dynamic evolves from one of direct control to one of collaboration and negotiation between human intention and algorithmic output.

The integration of AI tools into textile making introduces a power dynamic exceeding simple control; rather, it represents a complex interplay of influence affecting designer agency. This agency isn’t solely determined by the designer’s ability to direct the AI, but also by the AI’s capacity to suggest, modify, or even independently generate design elements. Consequently, designer agency can be enhanced through AI assistance, allowing for rapid iteration and exploration of novel ideas, or diminished if the designer feels constrained by the AI’s outputs or lacks the ability to effectively refine them. The degree to which AI tools augment or impede a designer’s creative control is therefore dependent on the specific tool, the designer’s skillset, and the nature of their interaction with the technology.

Relational agency, in the context of human-AI collaboration, refers to a designer’s perception of, and interaction with, the perceived agency of an AI tool; research indicates this is a key determinant of collaborative success. This isn’t simply about whether a designer believes an AI has agency, but how that perception influences their workflow and decision-making. Specifically, internal cognitive processes – including attribution of intent, expectation of outcomes, and assessment of responsibility – mediate the interaction. Designers who actively consider the AI’s ‘contribution’ and adapt their creative process accordingly demonstrate greater agency themselves, leading to more effective and satisfying collaborations. Understanding these cognitive mechanisms is crucial for designing AI tools that support, rather than diminish, designer agency and promote a balanced power dynamic.

Cultivating Self-Awareness Through Reflective Practice

Designers can employ self-reflection techniques, augmented by practices like design bookkeeping and maintaining a ‘Portfolio of Loose Ends’, to systematically analyze their engagements with AI tools. Design bookkeeping involves detailed recording of interactions – prompts used, outputs generated, and the designer’s immediate responses – providing a traceable history of the creative process. A ‘Portfolio of Loose Ends’ functions as a curated collection of incomplete ideas, discarded explorations, and unresolved questions arising from AI-assisted design, offering insight into areas where the technology presents challenges or sparks novel avenues for investigation. These methods facilitate critical examination beyond simply assessing usability; they allow designers to unpack their subjective experiences, identify cognitive biases, and understand how AI influences their creative decision-making.

Traditional usability testing assesses whether a user can achieve a specific task with an AI tool; however, techniques like self-reflection prioritize the designer’s internal experience during interaction. This approach moves beyond quantifiable metrics to explore qualitative data regarding the designer’s cognitive processes, emotional responses, and evolving mental models. The focus is not simply on what the designer accomplishes with AI, but how the interaction shapes their creative thinking, alters their workflow, and influences their perception of design agency. This emphasis on subjective experience allows for a nuanced understanding of the designer-AI relationship, revealing subtle shifts in practice that would be missed by objective performance measures alone.

Iterative self-reflection practices demonstrate a discernible impact of AI tools on established creative workflows and, critically, on the perceived autonomy of designers. Analysis of designer experiences reveals shifts in agency as AI handles tasks previously central to the creative process, prompting re-evaluation of skill sets and the definition of design expertise. This research confirms that a detailed understanding of designers’ internal cognitive processes-including their perceptions of control, creative ownership, and evolving problem-solving strategies-is essential for effectively integrating AI technologies into design practice and mitigating potential disruptions to creative work.

Towards More-Than-Human Collaboration in Digital Craftsmanship

Reflective practice, a cornerstone of professional development, reveals compelling opportunities for ‘More-Than-Human Collaboration’ within the evolving field of ‘Digital Craftsmanship’. This approach moves beyond viewing artificial intelligence as a simple instrument, instead recognizing its capacity for contributing unique perspectives and initiating novel pathways in creative endeavors. Through careful introspection on the design process – specifically, analyzing interactions with AI systems – practitioners gain valuable insights into how to leverage these technologies not just for automation, but for genuine co-creation. The resulting understanding fosters a dynamic where the designer’s expertise intertwines with the AI’s computational abilities, potentially yielding designs and textile creations that transcend the limitations of either agent working independently. This iterative cycle of action and reflection is proving essential to unlocking the full potential of these synergistic partnerships.

The evolving paradigm of digital craftsmanship increasingly frames artificial intelligence not as a passive instrument, but as an active collaborator possessing a discernible form of agency within creative workflows. This perspective challenges traditional design methodologies centered on control, instead emphasizing a reciprocal relationship where AI contributes unique generative capabilities and influences the direction of the creative process. Recognizing this agency necessitates a shift in how designers conceptualize their role; they become facilitators of a shared creative endeavor, responding to and building upon the contributions of the AI, rather than simply directing its output. This acknowledgment opens up avenues for exploring novel aesthetic possibilities and innovative design solutions born from the synergistic interplay between human intention and algorithmic generation.

The evolving relationship between designers and artificial intelligence within textile creation necessitates a shift from traditional control-based workflows to a collaborative synergy. This research indicates that true innovation arises not simply from employing AI as a tool, but from acknowledging its capacity as an active contributor with unique generative potential. Successfully fostering this ‘More-Than-Human Collaboration’ demands introspection; designers must cultivate awareness of their own cognitive processes – how they adapt, interpret AI’s output, and integrate it into their creative vision. By prioritizing this internal adjustment and embracing a flexible, responsive approach, designers can unlock novel avenues for artistic expression and fundamentally reshape the landscape of textile making.

The exploration of agency within the design process, as detailed in the paper, resonates deeply with Grace Hopper’s assertion: “It’s easier to ask forgiveness than it is to get permission.” This sentiment mirrors the need for designers to proactively adapt and experiment with emerging technologies like LLMs, rather than awaiting rigid guidelines. The paper highlights how conscious introspection and expectation management are vital for designers to retain creative control; a degree of ‘forgiveness’ – of initial imperfect outputs or unexpected AI behaviors – becomes necessary to navigate the collaborative dynamic. Ultimately, a willingness to iterate and learn from these interactions, even when deviating from pre-defined paths, fosters a more robust and empowered design practice.

The Road Ahead

The study of agency, as presented, reveals a recurring tension inherent in collaborative systems. Each new technological partner – and large language models are merely the latest iteration – introduces a subtle shift in the locus of control. The temptation lies in outsourcing cognitive load, yet this very act diminishes the capacity for nuanced judgment, for the ‘feel’ of a solution. Every new dependency is the hidden cost of freedom, a principle that extends beyond design and into the broader architecture of intelligence itself.

Future research should focus less on what these tools can do, and more on how they change the designer. Longitudinal studies tracking shifts in creative practice, and detailed introspection into the experience of collaboration, are essential. A critical examination of the metrics used to evaluate ‘creative success’ is also needed; are we optimizing for novelty, or for a deeper, more human resonance?

Ultimately, the question is not whether designers will be replaced by these technologies, but whether they will be subtly reshaped by them. The challenge lies in maintaining a reflexive awareness of this process, in understanding that the tools are not neutral, and that the structure of the collaboration inevitably dictates the outcome. The organism adapts, or it fades.

Original article: https://arxiv.org/pdf/2603.12018.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/

See also:

2026-03-13 08:19