The AI brand incident isn't a future risk. It's already the majority experience — and more review won't fix it.
There's a number in an IAB study that should have stopped the industry cold, and mostly didn't: when marketers were asked whether their current safeguards for AI-generated content were sufficient, only 6% said yes.
Six percent. In a profession that is otherwise very good at telling itself things are under control.
The rest of that study explains why the confidence is so low. An IAB study found that 70% of marketers have already experienced an AI incident — an output that went off-brand, a claim they couldn't substantiate, a hallucination that made it into something published. Forty percent had to pull ads as a result. And more than 90% said they want an independent way to check AI output before it ships.
Read those together and the picture is unambiguous. This isn't a fringe risk raised by skeptics on the sidelines. It's the majority experience of teams already shipping with AI, and they know their defenses don't hold. They're describing the present, not worrying about the future.
So why does it keep happening?
Because almost every team is solving the wrong half of the problem. Enormous effort has gone into adoption — better prompts, more tools, faster generation. In Jasper's 2026 report, 91% of teams now use AI in their work. Almost none of that effort has gone into control.
The implicit assumption is that if the model is good enough, the output will be safe. It won't, because the model doesn't know your brand. It knows language in general. It has never read your approved-claims list, your regulatory constraints, your do-not-say words. It is fluent, confident, and operating blind.
Review is the thing AI broke
The default safeguard is human review — catch the problems before they ship. But review is precisely what AI overwhelms. Jasper's 2026 report found cross-functional review friction, the legal and brand and compliance back-and-forth, rose 3.4x in a single year. You have industrialized production and left inspection manual.
The incidents aren't a sign that the people are careless. They're a sign the system is mathematically guaranteed to leak.
That is what the 6% is actually measuring. Not pessimism — arithmetic. Those marketers can see that the only thing standing between generated content and the public is a human being reading it, and they can see how much content there now is.
The fix is upstream, not downstream
The teams that get this under control do one thing differently. They stop treating "on-brand and compliant" as something you verify after generation, and start treating it as something the generation is grounded in. They give their tools a single, authoritative, machine-readable source of brand truth — voice, approved claims, rules, compliance constraints — so output is shaped correctly at the moment it is made, and the risky material is caught by construction rather than by luck.
Notice what that does to the review queue. It doesn't remove the humans; it changes what reaches them. Instead of inspecting everything on the chance that something is wrong, they handle the genuine edge cases. Governance stops being a brake and becomes a rail.
Seventy percent isn't a statistic about somebody else's bad quarter. It's the baseline. And 6% is the honest internal assessment of what is currently protecting you. The teams that move first from "we hope it's fine" to "we can prove it's fine" are the ones who will scale AI without scaling their incident count.
kbie is brand governance for the AI era — it turns your brand into a verified knowledge graph, so everything you and your AI tools publish stays on-brand, accurate, and safe to ship. → https://kbie.ai
Top comments (0)