Here's the thing about the AI gold rush that nobody in a marketing deck wants to admit: consumers aren't impressed. In fact, they're actively annoyed. A new report from WordPress VIP — the enterprise arm of Automattic's publishing empire — surveyed 2,000 people and found that 60% of U.S. consumers consider "AI" in brand messaging to be an immediate turnoff. Not neutral. Not indifferent. A turnoff.

Let that sink in while your competitors are busy adding "AI-powered" to every product description they own.

The Trust Deficit Is Real — and It's Getting Worse

The report, which pulled responses from 1,200 U.S. adults and 800 enterprise decision-makers and CMOs in April, paints a pretty uncomfortable picture for anyone betting their brand strategy on AI visibility alone. Eighty-six percent of consumers said they don't fully trust AI-generated answers and still want to track down the original source. That's not a niche behavior — that's basically everyone.

The stat that should haunt every content strategist: 42% of respondents said they trust AI-generated answers without clear attribution less than airline fees, impenetrable privacy policies, and medical bills. That's a murderers' row of things people viscerally hate. Congrats, unattributed AI content — you've found your peer group.

And if you needed one more gut punch, nearly three in four respondents said the internet feels less human than it did a decade ago. Which, honestly, tracks. When was the last time you read something online and thought, "a real person who cares about this wrote every word"?

So Why Is Everyone Still Optimizing for AI Search?

Because the traffic numbers are moving, and enterprise teams follow traffic numbers. Sixty percent of enterprise respondents reported that referrals from AI search engines and answer platforms grew over the past year. Seventy-four percent said AI discoverability is a "main or significant priority." So brands are chasing AI visibility even as their actual customers are developing a Pavlovian aversion to the word "AI." That's the tension nobody has a clean answer for yet.

WordPress VIP's CTO Brian Alvey framed the bind succinctly: "People used to build websites for other people. Now you have to build websites for AI agents acting on behalf of those people." He's not wrong. If an AI intermediary can't parse and cite your content, you effectively don't exist in an expanding slice of how people discover information. But if the tiny fraction of users who actually click through to your site land on something that feels synthetic and untrustworthy, they're gone.

You're now optimizing for two audiences with almost opposite preferences. Welcome to the current state of web publishing.

What Consumers Actually Want (It's Not Complicated)

The report offers a few signals that are genuinely useful if you're building something:

• Source attribution matters. Thirty-three percent of consumers said clicking through to an original source remains their top trust signal. Clear attribution isn't just ethical — it's a conversion lever.
• Open access over walled gardens. Eighty percent of respondents said web content should remain openly accessible rather than controlled by a handful of large platforms. That's a strong mandate against the slow crawl toward AI systems that answer everything without ever sending users anywhere.
• Human voice is a differentiator now. If most content is getting flattened into AI-generated beige, writing that sounds like it came from an actual human with a perspective and a pulse is increasingly valuable. That's not a soft, fuzzy marketing claim — that's supply and demand.

The Obvious Caveat About Who's Asking

It's worth noting that this report comes from WordPress VIP, whose entire business model depends on brands continuing to invest in their own websites and content infrastructure rather than ceding everything to AI intermediaries. Their interest in an "open web" narrative is genuine — Automattic has backed open-source projects and open protocols like ActivityPub for years — but it's also conveniently aligned with their revenue model. Take the framing with the appropriate grain of salt, even if the underlying survey data holds up.

None of that makes the findings wrong. It just means you should verify the methodology before stapling these numbers to a board presentation.

The Bottom Line for People Actually Building Things

The companies that figure out how to be legible to AI systems and compelling to actual humans aren't going to do it by plastering "AI-powered" across their homepage. They're going to do it by producing content with genuine attribution, clear sourcing, and a voice that doesn't read like it was generated by averaging the internet. The brands treating AI as a label rather than a capability are already paying for it in consumer trust — they just haven't noticed the bill yet.

Sixty percent think "AI" in your messaging is a red flag. The other forty percent probably aren't paying close enough attention yet.