Smashing Magazine: no, people don't want more AI in their life

Vitaly Friedman, editor-in-chief of Smashing Magazine, challenges an assumption that runs through most product roadmaps in 2026: that users are eager for AI features. The data, he argues, tells a different story.

The problems he identifies are practical. AI features frequently arrive as disconnected add-ons that disrupt existing workflows rather than fitting into them. Users spend time verifying outputs, catching hallucinations that require correction. When AI amplifies organizational shortcomings rather than addressing them, it creates more problems than it solves. Among the general public, AI implementation is also associated with job replacement — a perception that creates resistance before a product is even opened.

Friedman distinguishes between AI that asks for attention and AI that disappears into the background. What most users actually want, he argues, is AI that handles tedious tasks without requiring them to notice it. The best implementations he describes as “AI-second”: subtle, deeply integrated, and supportive rather than prominent and disruptive. The outcome users respond well to is one where AI removes friction rather than adding a new thing to manage. Low adoption and retention rates for AI features are, in his reading, the predictable result of skipping this question.

The article is most useful for product designers and UX leads working on AI-adjacent features who need a counterweight to the default assumption that more AI equals better product. Friedman does not argue against AI in products; he argues against AI that was shipped without asking whether users need it in that particular form. The distinction matters in practice: the same capability delivered through a persistent chat panel versus an invisible background process can produce very different user responses, regardless of what the AI is actually doing.

For designers in organizations where the drive to ship AI capabilities is strong, the piece provides a frame and some supporting evidence for the internal conversation about restraint. It is short and direct, grounded in observable adoption patterns rather than speculation about where the technology is heading.