There's a moment most design leaders recognize. A piece of AI-generated work gets shared in a review — image, video, a visual concept — and the room accepts it. Not because it's good. Because it's fast, it's close enough, and no one in the room has the reference point to push back.
It's happening more than anyone is admitting.
The State of AI Design 2026 report found that 91% of designers now use AI in their work at least weekly, up from 54% just a year ago. The average toolstack has more than doubled. Half of designers have shipped code directly to production. By every measure of speed and volume, the tools are delivering.
The same report also found that 62% of designers cite inconsistent or unreliable output as their biggest challenge. The tools are fast. The output isn't always good. And the gap between those two facts is where the real problem lives.
The gap
AI didn't lower the craft bar. It made the bar easier to ignore.
When execution was expensive, the cost of bad work was obvious. A poor visual direction meant wasted production budget, missed timelines, a conversation with a client. The constraint created accountability. Now that execution is cheap, that friction is gone, and judgment has to replace it — the ability to look at output and know whether it's working, not just whether it's finished.
That accountability has to live somewhere, and right now it often doesn't.
The approval problem
Some designers are accepting AI output because the output is genuinely strong and their judgment is calibrated to recognize it. That's the use case the tools were built for. But some are accepting it because they don't have a clear reference point for where good ends and adequate begins, and because no one above them has set that reference point either. When that happens, adequate ships. Adequate gets approved by account teams who don't know what they're looking at. Adequate lands in front of clients who mistake fast turnaround for craft.
A significant portion of the AI-generated design work being approved right now is below the standard it appears to meet. Just slightly off, in texture, in coherence, in the kind of fine-grain quality that separates work that builds brand trust from work that quietly erodes it.
The people approving it often can't tell. Evaluating visual quality at a professional level requires training and reference. Most account leads, most clients, most cross-functional partners don't have that. They're evaluating on feel and on speed, and AI output has become fluent enough to pass both tests. Passing feel and speed is not the same as passing craft.
"Fast and good are not the same thing. They can coexist, but one doesn't guarantee the other."
The accountability question
The question this creates for design leaders isn't about which tools to use or how to upskill teams on prompting. It's simpler and harder than that: who in your organization is responsible for knowing when the work isn't good enough? Who has the authority and the developed eye to say so, when everyone else in the room has already moved on?
Those questions deserve a clearer answer than most organizations currently have.
The speed trap
The tools have removed the natural friction that once forced quality conversations. Before, the cost of iterating was high enough that people argued about whether a direction was right before committing. Now the cost is close to zero, which is a genuine creative advantage, but it removes the moment where someone has to defend the choice. Work flows through faster. The reckoning, if there is one, comes later.
For cross-functional leaders, the CMOs, CPOs, and account leads championing AI as a speed play, this is worth sitting with. Fast and good are not the same thing. They can coexist, but one doesn't guarantee the other. The moment you measure design output by how quickly it was produced, you've made a decision about what you value. And the work will reflect that decision over time, in ways that aren't always visible until they've already done their damage.
Speed is not a standard. Someone still has to set one.
Looking for more thought leadership?