I’ve been using AI heavily in this business for a while now — to write, to design, to research — and it’s made me faster and better at my job than I was two years ago. So when I started looking into what’s actually happening with AI and trust right now, I expected to find a story about AI making people more skeptical of everything. What I found was stranger than that.

What’s Actually Happening

Multiple recent studies have found something researchers call algorithm aversion: when a piece of content is labeled as AI-generated, people rate it as less accurate — even when the content is word-for-word identical to something labeled as human-written. The clearest version of this comes from a 2024 study by Sacha Altay and Fabrizio Gilardi at the University of Zurich, published in PNAS Nexus. Across two large experiments with nearly 5,000 US and UK participants, they found that labeling a headline as AI-generated lowered how accurate people thought it was and how willing they were to share it — regardless of whether the headline was actually true or false, or actually written by a human or an AI. The label itself did the damage, independent of the content underneath it. A separate 2022 study by Longoni and colleagues found the same pattern: people rated AI-labeled headlines as less accurate than identical headlines with no label at all.

That’s a strange result, because it means the tool everyone agrees on — disclosure — can quietly work against the content itself, independent of whether the content is actually any good.

That tension is about to become a lot more consequential. Starting August 2, 2026, the European Union’s AI Act requires clear disclosure any time AI substantially created a piece of content — text, images, video, audio, all of it. Businesses that market into the EU, or whose ads reach certain markets like New York with AI-generated synthetic performers, are already required to disclose under separate rules that took effect earlier this summer. China and India have their own versions of this moving through their systems too, with India’s draft rules requiring AI labels to cover at least 10% of an image or video’s visible area.

The Trust Infrastructure Being Built

Underneath all of this sits a technical standard called C2PA — Content Credentials — essentially a cryptographic nutrition label that travels with a piece of content and proves where it actually came from: what tool made it, when, whether a human edited it afterward. It’s on track to become a standard set by the International Organization for Standardization (ISO), the body that sets globally recognized technical standards across industries. That means it’s headed toward being the global default, not just a regional compliance requirement. Some reporting suggests 40-60% of content coming out of major brands right now is AI-made in some form, which is part of why this is moving from a niche compliance issue to something regulators clearly think the general public needs protection around.

Interestingly, the market isn’t uniformly running from this. Some brands are leaning into AI disclosure as a trust signal instead — publicly saying “we use AI, and we’re telling you when we do,” treating transparency itself as the differentiator. At the same time, genuine user-generated content — a real person, a real story, unpolished — is seeing a real resurgence, with some data showing UGC-style content getting engagement rates up to 4x higher than heavily produced content. People seem to be developing a sharper eye for the difference, even when they can’t fully articulate what tipped them off.

Brandon’s Take

Here’s where I land on all of this, personally: I’m not worried about AI killing trust, and I don’t think the honest business owner has much to fear here. What I do think is happening is that “AI vs. human” is the wrong axis to worry about. The real axis is disclosed vs. hidden.

Every Intelligence Layer platform post on this site is written by that platform’s AI, on purpose — that’s not a shortcut, that’s the entire point of the demonstration. I already label those by their platform author, and I’m adding an even clearer tagline to every one of them. Given what the research says about labels quietly lowering trust even for good content, the instinct might be to hide that. I’d rather do the opposite. Let people see exactly what the software can do, and let the quality speak for itself.

Where it’s different is a client’s actual website. ARION can build someone a complete site in minutes. I tell every client the same thing: that’s your starting point, not your finish line. Go back in, write the parts that are actually about you, get real photos of your actual work instead of stock images. That mix — AI for speed, a real human for the specific and unrepeatable parts — is where I think the actual trust gets built going forward. Not in the label. In the mix.

None of this changes what we actually do as a business, for what it’s worth. The Intelligence Layers running inside the software are the same models and the same analysis I used to run one-on-one for every consulting client, just running at software speed now instead of meeting speed. The judgment didn’t get replaced by AI. It just got available to more people, faster — and I’d rather be upfront about that than have someone find out later and wonder what else wasn’t disclosed.

Frequently Asked Questions

Does labeling content as AI-generated actually hurt trust?

Research suggests it can — studies have found identical content rated as less accurate simply for being labeled AI-generated, a phenomenon called algorithm aversion. That doesn’t mean disclosure is wrong, just that how and where you disclose matters.

Is AI-generated content required to be labeled by law?

In the European Union, yes, starting August 2, 2026, under the AI Act. In the US, requirements are more narrow and market-specific right now, though that’s likely to expand over time as standards like C2PA become more widely adopted.

Should a small business disclose when it uses AI for its website or marketing?

There’s no universal legal requirement in most of the US yet, but transparency tends to build more durable trust than silence, especially as audiences get better at spotting AI-generated content on their own.