Lead image for the Muzli essay The productisation of taste

Can AI Learn Taste, or Only Reflect Ours Back?

An essay asking whether AI can learn taste, and why its everyday defaults may shape what creative work we come to expect.

Strategy

1 min read

Can Machines Learn Taste, or Only Imitate It?

This essay argues that the common claim that AI lacks taste rests on an unexamined assumption: that machines can get better at making things while staying hopeless at judging them. The author suggests taste is not a mysterious faculty but a learnable skill of knowing which conventions to follow, which to break, and what each choice achieves. A 2026 study comparing AI-written passages with trained writers is used to show that machine output can already meet aesthetic standards people associate with taste.

For brand builders and designers, the more pressing concern is influence rather than quality. The piece warns that when AI systems keep returning to safe, approved choices, those defaults can quietly become the benchmark for good work and narrow what teams consider possible. It also sketches a more hopeful future in which accessible creative tools let more people give form to ideas. Anyone who makes choices that shape audience expectations, from identity systems to campaign aesthetics, should read it.

Read Full Story

Source: muz.li · Originally published

Published: Updated: