Taste or Discernment

In the context of AI and design, one common talking point lately seems to be the importance of "taste". The argument here is supposed to be that since it is supposed to be so much easier and faster to produce outputs, deciding which ones are the ones to keep is an important differentiator for designers in the age of AI.

First of all, reducing this to a matter of taste highlights the all too problematic notion that design should be judged based on something as subjective as taste. It puts us back to discussions over whether a button should be blue because an important stakeholder simply does not like blue.

I believe a more apt description is discernment, a delicate discrimination of something, the trait of judging wisely and objectively. And this requires a sufficiently deep understanding of the respective context.

Where taste may evaluate something at a superficial level, based on aesthetic qualities, the idea of discernment goes beyond this limited scope. Is something built right? Is it the right thing to build to begin with? Does it actually meet a user's needs? Is this truly the most important benefit for said user? Is this the best solution to provide this benefit?

Notice that even the evaluation of something being the "best" in these questions requires a lot of nuance and careful definition. Cost-effective? Well-integrated into existing tools and workflows? Easily scalable? Accessible? Inclusive? Sustainable? Various stakeholders will have differing evaluation criteria for these questions, as do users. Understanding these criteria is of the utmost importance to evaluate solutions - whether produced by AI, humans, or ... well, these days, who knows.

Thus, I think it is not a matter of taste that differentiates designers in such scenarios. It is contextual awareness and understanding. It is building bridges between user needs, business context, and technological realities. Which looks a lot like desirability, viability, and feasibility. Which isn't actually a new requirement for the age of AI, but something any good designer should be familiar with.