The conversation about AI in design tends toward two extremes. On one side, the idea that AI will render designers obsolete, reducing craft to a prompt and a generation. On the other, the dismissal of AI tools as toys, capable of approximating style but incapable of anything that actually matters.

Neither position is particularly useful, and neither matches what is actually happening in practice. The more interesting question is not whether AI will replace designers but how designers who use AI well will work differently from those who do not.

What AI is actually good at

AI tools currently excel at a specific category of tasks: generating variations, exploring a solution space quickly, handling the mechanical parts of production. Resizing and adapting assets across breakpoints. Drafting copy that can be refined. Generating a dozen layout directions from a brief so that a designer can spend their time reacting and editing rather than starting from a blank canvas.

These are not trivial things. A significant portion of design work at every level is exactly this kind of production and iteration. Offloading it to tools that can do it faster, even imperfectly, genuinely changes how time gets spent.

What still requires a person

AI does not understand context in the way that matters for design. It can recognize patterns in its training data and generate outputs that fit those patterns, but it cannot understand the specific circumstances of a user sitting in a hospital waiting room trying to update their insurance information, or the cultural connotations of a visual choice for a particular audience, or whether the tone of an error message will feel condescending to the people who will read it.

These judgments require understanding. They require knowledge of the specific people the product serves, the specific context in which it will be used, the specific history and values of the organization building it. AI can assist with the execution of a judgment once it has been made. It cannot reliably make the judgment itself.

That is not a temporary limitation waiting to be solved. It reflects something real about what design work is at its core.

How the role is shifting

What changes when AI handles more of the production work is not that designers become unnecessary. It is that the ratio of time spent thinking versus executing shifts. Designers who adapt to this well spend more time earlier in a project, doing the work that grounds everything else: understanding users, defining problems, making decisions about what the product should and should not try to do.

This is not a new direction for the profession. It is an acceleration of a direction the field has been moving in for years. The designers who have always prioritized strategy and research over pure execution find the shift more natural. Those who have built their practice primarily around production skill face a steeper adjustment.

The question of authenticity

One concern that comes up regularly in design communities is whether AI-assisted work is authentic, whether there is something lost when a visual was generated rather than drawn, a layout derived rather than reasoned through.

This is a real question worth taking seriously, but it is also not a new one. Designers have always used tools that do work on their behalf: grids, templates, component libraries, stock photography. The question has never been whether tools are involved but whether judgment is. Judgment about what to make, for whom, in service of what goal. That question does not go away with more capable tools. It becomes more important.