AI is now a routine participant in client design reviews. A veteran designer with 15 years of client history recently discovered their client had fed a completed website design to ChatGPT for a second opinion. The client meant no harm. That is precisely what makes this a structural shift, not an isolated incident.
The threat is not bad AI feedback. Bad feedback is easy to dismiss. The threat is plausible feedback, the kind that requires a designer to spend billable time unpacking synthetic recommendations generated without knowledge of the actual business context, budget, timeline, or user base. The article traces this dynamic carefully, including why controlled AI critique inside a designer's own workflow is functionally different from a client running unsupervised prompts and presenting the output as a competing brief.
The piece lands on four concrete responses: don't treat it as a personal challenge, ask to see the original prompts, evaluate the ideas on merit regardless of source, and accept that the designer's role is shifting from producer toward curator and translator. The argument that scarcity of human judgment increases as AI suggestions multiply is the thread worth following into the full read.
[READ ORIGINAL →]