For decades, designers have approached new technologies primarily as tools—instruments that extend our capabilities but remain distinctly separate from the products we create. From the earliest CAD programs to modern prototyping software, technology has served as an assistant in the design process, not as a fundamental component of the design itself. However, as artificial intelligence rapidly evolves and becomes more deeply integrated into products, a profound shift occurs: AI is not just transforming, but revolutionizing from a mere tool into a material—something designers must learn to shape, mold, and incorporate as an essential element of their work. This means that AI is not just a tool that assists in the design process, but a material designers can use to create more adaptive, personalized products capable of evolving alongside their users.
This article is the first in a four-part series exploring the concept of "AI as a material" in product design. The series will unfold as follows:
Part 1: AI as a Material: A New Design Paradigm - Introducing the concept and exploring AI's unique material properties
Part 2: Working with AI as a Material: New Design Processes - Examining how design methodologies must evolve to work with this new medium
Part 3: The Designer's Role in an AI Material World - Investigating how designers' skills and responsibilities transform when working with AI materials
Part 4: The Future and Ethics of AI as a Material - Considers ethical implications and future directions as AI materials evolve. It includes issues such as data privacy, algorithmic bias, and the potential for AI to replace human creativity in design processes. By addressing these ethical considerations, designers can ensure that their AI integration is technically impressive and genuinely beneficial, ethical, and human-centered.
My journey into this topic stems from a fascination with how our metaphors for AI have evolved over time—from chat bot, assistant, co-pilot and agent to potentially a material and creative collaborator. These metaphors aren't merely semantic distinctions; they fundamentally shape how we approach design challenges and envision possibilities. By exploring AI as a material, I hope to develop a deeper understanding of what this means for product design—not just how we use AI tools to design, but how we create experiences and products where AI itself is a core component, potentially even an agentic one that collaborates in the creative process.
This perspective shift is more than semantic. When we view AI as a material rather than a tool, we fundamentally change how we approach design challenges, opening new possibilities while imposing different constraints. Just as a furniture designer must understand the properties of wood—its grain, strength, and flexibility—the product designer must develop an intuitive understanding of AI's unique characteristics, limitations, and potential.
This first article dives into the emerging paradigm of AI as a design material, highlighting its properties and how it differs from traditional materials. By treating AI not just as something that aids in design but as something we design with, we can develop products that are more adaptive, personalized, and capable of evolving alongside their users. For instance, AI can be used to design smart home products that learn and adapt to users' habits or create personalized shopping experiences online.
Why Understanding AI as a Material Matters for Designers
As AI becomes increasingly integrated into products and experiences, designers who understand how to work with it as a material will have significant advantages creatively and professionally. These practical benefits should motivate designers to embrace AI as a material in their work.
Improved Design Outcomes
Viewing AI as a material enables designers to create more sophisticated, nuanced products that adapt to individual needs and contexts. Rather than designing static experiences that quickly become obsolete, designers can create products that evolve with users' behaviors and preferences, extending their relevance and utility. This leads to experiences that feel more intuitive, personal, and responsive than would be possible with traditional design approaches.
Competitive Professional Advantage
As organizations increasingly integrate AI into their products, the need for designers who can think beyond surface-level implementation is growing. Designers who understand AI's unique material properties can bridge the gap between technical capabilities and user needs, positioning themselves as valuable strategic partners rather than interface creators. This understanding opens doors to new roles and opportunities that may not be available to designers who view AI merely as a tool.
Enhanced Creative Possibilities
Working with AI as a material expands the palette of what's possible in design. Just as the introduction of new physical materials throughout history—from plastics to touchscreens—enabled entirely new categories of products, AI as a material allows designers to create experiences that would be impossible through other means. This includes products that learn, adapt, anticipate needs, and even contribute creatively to solving problems alongside users.
By developing a deeper understanding of AI as a design material, designers can position themselves at the forefront of this transition, helping shape how AI is integrated into products in genuinely beneficial, ethical, and human-centered ways rather than merely technically impressive ones.
The Evolution From Tool to Material
What differentiates a tool from a material in design? Tools extend human capabilities but remain separate from the finished product. Materials, conversely, become integral components of what we create—they have properties that must be respected, constraints that must be accommodated, and possibilities that can be exploited. Throughout design history, we've witnessed technologies transition from tools to materials as they become more accessible and malleable.
Electricity provides a compelling historical parallel. Initially approached as a tool for powering machinery, electricity eventually became a design material in its own right. Lighting designers don't merely use electricity as a tool; they design with it, shaping its qualities, intensities, and behaviors. Similarly, early digital interfaces were first created using software tools, but eventually, interactivity itself became a material that designers learned to manipulate directly.
AI is following this same trajectory, moving from an external tool that helps generate designs to an embedded material with properties that designers must understand and work with intentionally. This transition is not just a shift in perspective; it's a call for a new vocabulary, methodologies, and ways of thinking about the design process. It's an urgent and essential step in the evolution of design that designers must embrace to stay at the forefront of their field.
The Material Properties of AI
Just as physical materials have properties like hardness, flexibility, or opacity, AI, as a material, has its own characteristic properties that designers must learn to understand:
Adaptivity: Unlike static materials, AI can change its behavior based on input and usage patterns. This quality makes it more comparable to living materials than traditional ones. In products, this manifests as personalized music recommendations that evolve with listening habits, email clients that prioritize messages based on response patterns, or navigation apps that learn commute preferences and suggest faster routes based on your typical schedule.
Generativity: AI can produce outputs that weren't explicitly programmed, creating emergent behaviors and possibilities. This appears in creative tools that can extend a designer's sketch into multiple finished concepts, writing assistants that can develop a rough outline into a complete draft, or game environments that generate new challenges tailored to a player's skill level and preferences.
Memory: AI systems can retain information from previous interactions, allowing for experiences that evolve over time. This property enables smart home systems that remember which lighting scenes you prefer at different times of day, language learning apps that focus on vocabulary you previously struggled with, or health trackers that recognize patterns in your wellness data across weeks and months.
Opacity/Transparency Spectrum: Some AI systems are more explainable than others, affecting how visible their "inner workings" are to both designers and users. This property is evident in credit scoring applications that must explain their decisions, medical diagnostic tools that show confidence levels and reasoning paths, or content recommendation systems that allow users to adjust the factors influencing their feed.
Embedded Values: Every AI system reflects the data it was trained on, including any biases or assumptions present in that data. This manifests in hiring tools that might perpetuate gender biases without careful design, facial recognition systems that may perform differently across demographics, or search algorithms that could amplify existing societal viewpoints without deliberate balancing.
Temporal Dimensions: AI behaviors can change over time as new data is incorporated, meaning designers must consider not just current states but future evolutions. This appears in customer service chatbots that incorporate new product information without redesign, content moderation systems that adapt to evolving cultural contexts, or autonomous vehicles that improve their driving capabilities through fleet learning.
These properties make AI fundamentally different from traditional design materials, requiring new approaches and considerations (Kumar & Schmitt, 2025).
To better understand how the material perspective applies to AI, let's consider a conceptual example: a smart home system. Traditional approaches treat AI as a tool that powers a product's features—voice recognition, pattern detection, and automated responses. But viewing AI as a material means recognizing that the system's learning capabilities, adaptability, and generative potential are themselves elements that can be shaped and designed.
When treated as a material, the AI within a smart home becomes something that designers would explicitly mold, determining how quickly it adapts to user patterns, balances predictability with suggestion, and makes its decision-making processes transparent to residents. These qualities wouldn't just be features implemented with AI; they would be manifestations of AI's material properties that have been intentionally shaped through design decisions.
This conceptual shift encourages designers to think beyond implementing isolated "smart" features and instead focus on crafting the fundamental properties of intelligence—its responsiveness, memory depth, adaptation rate, and how it communicates its learning to users. This hypothetical example illustrates how the material perspective transforms our approach to AI integration in everyday products.
Physical vs. Digital Material Properties
Traditional design materials exist in physical space, subject to physics and material science laws. Their properties—weight, texture, conductivity, flexibility—are measurable and largely stable. AI, as a digital material, operates under different constraints. Its properties revolve around information processing, pattern recognition, prediction, and adaptation.
However, AI doesn't exist purely in the digital realm. It interfaces with the physical world through sensors, displays, speakers, and actuators. This duality creates a unique hybrid material that spans physical and digital domains, requiring designers to consider both dimensions simultaneously.
The "Grain" of AI
Woodworkers understand that wood has a grain—a natural direction in which it splits easily. Working against the grain creates resistance and potentially undesirable results. Similarly, AI has natural tendencies and resistance points:
AI excels at pattern recognition but struggles with causal reasoning
It can generate variations efficiently, but may be less effective at true novelty
It learns effectively from abundant, clear examples but struggles with rare edge cases
It can process vast amounts of data, but may miss contextual nuances that humans grasp intuitively
Understanding this "grain" of AI helps designers work with its natural tendencies rather than against them, just as a skilled woodworker doesn't fight against the natural grain of their material but instead designs with it in mind, AI designers shouldn't try to force AI to perform tasks against its natural tendencies. For example, making an AI system perform nuanced ethical reasoning based purely on pattern recognition would be like trying to carve intricate details against the grain of wood—possible with exceptional effort, but likely to produce suboptimal results and be prone to unexpected failures.
Constraints and Limitations
Every material has constraints, and AI is no exception:
Computational Requirements: Advanced AI systems often demand significant processing power and memory, affecting device size, cost, and energy consumption.
Data Dependencies: AI's capabilities are bounded by the data it has been trained on, creating potential blind spots or biases.
Evolving Standards: Unlike established materials with industry standards, AI lacks consistent benchmarks and certifications. This forces designers to develop proprietary guidelines, complicates cross-team communication, and creates risk when designing for capabilities that may become regulated or obsolete.
Explainability Challenges: More complex AI systems may function as "black boxes," making it difficult to understandexactly how they arrive at certain outputs.
Energy Consumption: Training and running AI systems can require substantial energy resources, raising sustainability concerns.
These constraints don't diminish AI's value as a design material, but they shape how it can be effectively incorporated into products.
New Design Possibilities
Despite these constraints, AI's unique material properties enable design possibilities that would be impossible with traditional materials:
Personalization at Scale: Products that truly adapt to individual users without requiring manual customization.
Ambient Intelligence: Environments that subtly respond to human needs without explicit commands.
Evolving Experiences: Products that improve and change over time rather than degrade with use.
Anticipatory Design: Interfaces that predict user needs rather than simply responding to explicit inputs.
Collaborative Creation: Products that actively participate in the creative process alongside their users.
These possibilities aren't merely AI-enabled features but fundamental characteristics of products where AI serves as a core material element (Dovetail, 2023).
Is "Material" the Right Metaphor?
Some designers may question whether viewing AI as a material is the most appropriate metaphor. After all, AI lacks the physical tangibility of wood or metal, and its capabilities continually evolve in ways traditional materials do not. Critics might argue that AI is better understood as a system, service, or collaborator rather than a material.
These concerns are valid, but the material metaphor offers unique advantages over alternative frameworks. Unlike systems or services frameworks, which tend to emphasize technical implementation, the material perspective focuses designers on AI's experiential qualities and inherent properties. It encourages designers to consider how users interact and experience AI's behaviors rather than getting lost in technical specifications or algorithmic details.
Furthermore, while the collaborator metaphor (which we'll explore in depth in Part 3) captures important aspects of advanced AI interactions, the material metaphor provides a more accessible starting point for designers. It connects to familiar design thinking approaches while introducing new considerations specific to AI. The material perspective also maintains appropriate creative agency with the designer, positioning AI as something to be shaped rather than a semi-autonomous partner that might have its own agenda.
Rather than seeing these metaphors competing, we can view them as complementary perspectives that become relevant at different stages of AI sophistication and for different types of AI applications. The material metaphor offers a particularlyvaluable framework for designers seeking to integrate AI's capabilities into cohesive, user-centered experiences.
Looking Ahead to Working with AI Materials
The shift toward understanding AI as a design material rather than merely a tool represents a fundamental change in how we approach product design. By recognizing AI's unique material properties—its adaptivity, generativity, memory, and embedded values—designers can create entirely new categories of products and experiences that evolve, adapt, and collaborate with users in unprecedented ways.
As we'll explore in the next article of this series, this new perspective demands new design processes and methodologies. Designers will need to move beyond static wireframes and mockups to prototype with "living" AI materials that learn and adapt. They'll need new approaches to testing that account for emergent behaviors rather than fixed interaction patterns. And they'll need frameworks for ethical considerations that go beyond traditional usability concerns.
Part 2 will delve into these practical aspects of working with AI as a material, introducing concepts like "prompt engineering as material manipulation," "training data as material preparation," and new collaborative approaches that bridge design and data science disciplines. We'll examine how design teams are already developing new processes to effectively shape AI materials into thoughtful, human-centered products and experiences.
This journey from tool to material represents not just a shift in terminology but an evolution in how we conceptualize our relationship with technology—one that opens exciting new possibilities for designers ready to master this powerful new material.
Dovetail. (2023, July 22). How AI is supercharging product design (2024 guide). https://dovetail.com/product-development/ai-in-product-design/
Kumar, V., & Schmitt, G. (2025). AI as a co-creator and a design material: Transforming the design process. Design Studies. https://www.sciencedirect.com/science/article/pii/S0142694X25000158