AI Is Getting Better at “Taste”—But Humans Still Define What Good Looks Like

AI Is Getting Better at “Taste”—But Humans Still Define What Good Looks Like

Fast Company argues that as AI-generated design improves, “taste” is becoming the next major frontier for creative AI. AI can now produce technically correct websites, images and other creative work, but expert designers can still recognize patterns that feel generic or derivative—such as repetitive layouts, floating dashboards and excessive gradients. The challenge is no longer simply getting AI to make something correct; it is teaching it to make something that feels distinctive and genuinely good.

The problem is that taste is subjective and domain-specific. Companies such as Figma and Krea are therefore collecting curated examples from professional designers, architects, filmmakers and other experts to post-train their models. Krea, for example, has gathered specialized references covering film styles, printing techniques, architectural materials and other professional visual vocabulary, while its collaboration with architecture firm Henning Larsen provides expert feedback to improve its outputs.

This means that making AI “tasteful” actually requires a huge amount of human judgment. Companies such as Surge use professional writers and creatives to compare and evaluate AI outputs, while Contra Labs connects its network of creative professionals with AI developers for expert evaluations. These human judgments become training data that teaches models what experts consider good, useful, varied or distinctive. In effect, AI's apparent taste is being built from thousands of human aesthetic decisions.

The broader takeaway is that AI may automate creative production without eliminating the importance of creative expertise. As generating competent work becomes easier, the scarce skill shifts toward knowing what deserves to be made, what feels original and why one version is better than another. AI can increasingly produce the output, but today's best systems still depend heavily on humans to define quality—and that suggests that taste, judgment and creative direction could become even more valuable in an AI-driven creative economy.

About the author

TOOLHUNT

Effortlessly find the right tools for the job.

TOOLHUNT

Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to TOOLHUNT.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.