AI has entered the design chat(gpt). Not only are software companies scrambling to implement AI features into their products, but tools like ChatGPT, Bard, DALL-E, Stable Diffusion, and others are placing the power of AI in the hands of the masses. Entirely new products, such as Adobe Firefly, are emerging and essentially replacing stock photography (and Adobe Stock). This shift feels reminiscent of the mid-90s when internet browser updates were released weekly, and the capabilities of designers seemed to grow exponentially. In the case of AI, this is certainly true.

We see people posting examples of AI creations in different media, such as interface designs, art, prose and video entertainment. These creations are driven by designers providing prompts and parameters to get their desired result from an AI engine. AI-based design toolkits are already integrated into many of our design apps and workflows. We’re already seen Figma plug-ins and platforms like Stable Diffusion being used to create high fidelity design work that sometimes surpass what the human operator can do. Is this just skilled execution or is it design mastery that gives us a great result? As we think about UX design and the idea of co-creating with AI, we have to wonder: Does design mastery matter? If so, where is it–Is it in the prompt, the execution, or somewhere else?

Redefining design mastery

As designers, we want to believe that our work embodies an improvisational blend of art and problem-solving—an essence that AI can never truly replicate. A master designer possesses a deep understanding of their craft and can decide when to honor the rules and when to break them for effect. We believe this combination of artistic flair and calculated decision-making is something that sets us apart from AI. And, our human touch will always be an irreplaceable component of the design process. But, what does it mean when AI contributions become more impressive, more helpful and more difficult to distinguish from human contributions each day? It means that the placement and degree of that human touch is shifting.

As AI tools have been introduced, some have been labeled “junior” creators or “assistants” that may provide robust, albeit imperfect, contributions that the designer must ultimately decide to use or ignore. This position is a comfortable one to adopt, as it feels more like an extension of a designer’s capabilities without any loss of control. We’re still calling the shots, even if we may be prompting the AI along the way to make things more efficient or diverse.

As AI expands the designer’s capabilities, we may find ourselves working beyond our normal boundaries and going into unexplored territories. A visual designer might prompt an AI writer for headlines, while a user researcher could be crafting high fidelity prototypes, courtesy of an AI design tool. Regardless of the novel space being explored, these AI tools will accelerate a designer’s journey beyond their comfort zone into a place where they are less competent and more receptive to suggestions and adopting ideas. In such instances, it’s beneficial to view design mastery not as control or dominance, but as the achievement of a design goal, utilizing any tool that contributes to success. Consequently, the output of AI may be directly woven into the final product. This approach truly augments one’s capabilities, enabling designers to create in new areas made possible by AI.

Design mastery, still up to us humans.

If you replace “AI” with “teammate,” everything we’ve mentioned would sound like business as usual for the typical designer. We’ve all collaborated with team members who “assisted” by conducting audits, analyzing design patterns, and supporting decision-makers. We’ve all worked side by side, dividing the tasks, sharing responsibilities, and ensuring everything comes together seamlessly in the end. In these aspects, working with an AI feels natural, albeit with a different interface and a heightened need to double-check the output.

We need to undergo a dramatic shift when it comes to creating work using AI that we normally wouldn’t attempt ourselves. It’s in those moments where we let AI run wild that things can go off the design rails. We’ve used ChatGPT to analyze survey data, and while the first few prompts yielded impressive results, that quickly became unreliable. As we pushed the AI to manipulate the data for higher-level insights, we noticed new data had magically been inserted into the dataset. When questioned, ChatGPT apologized for adding its own data and removed it. While taking accountability for errors is admirable, it was unsettling to see just how easily the integrity of the work could be compromised. Had we not been vigilantly monitoring its progress, we could have easily strayed down a path where our design decisions would end up being compromised by the AI’s work. The moral of the story is that while it’s tempting to sit back and let AI “do the work,” a sense of responsibility and design mastery are still required. The onus shifts to the management of design, the strategic thinking guiding design and the pivotal design decisions that shape the final form. AI lacks the intelligence and touch that humans possess when it comes to ensuring design work is done right and feels appropriate to the moments we create when our users interface with the work.