Someone asked me directly a few weeks ago, kind of testing me: “did AI write this newsletter?” I told her the truth, no hedging. Parts of it, yes. I write the ideas, the opinions, the actual point of each issue. AI helps me get from a rough idea to a clean draft faster, the same way I’ve always used an editor, a whiteboard, or a stubborn friend who tells me when something isn’t working.
She wasn’t upset. She said, “Hm. Well okay, I guess, as long as it’s actually useful and it’s actually you.” That’s the whole debate, distilled into one sentence by someone who clearly doesn’t need a newsletter to explain trust to her.
Yes, AI helped write this. Here’s why that’s fine, and where the line is.
There seems to be a fear out there that if people find out AI helped with something, they’ll feel lied to. I don’t think that’s actually true, and the data doesn’t back it up either. What erodes trust isn’t the tool. It’s getting something generic, wrong, or hollow and suspecting a machine is the reason, which is really just a trust problem wearing an AI costume.
Flip it around. If a piece of content genuinely helps someone, answers their actual question, saves them actual time, tells them something actually true, most people don’t care what tool was involved in drafting it. They care whether it was worth their time to read. That’s always been the bar. AI didn’t lower it or raise it, it just made it more obvious which businesses were meeting it and which ones were coasting on effort alone.
Here’s where the actual line sits, and it’s not about disclosure percentages or a little badge that says “AI-assisted.” The line is whether a real person stood behind the final version. Did someone with actual judgment, actual knowledge of the business, actual stake in getting it right, look at this before it went out and decide it was true and worth sending? If yes, the AI involvement is a production detail, not a deception. If no, if it went straight from a prompt to a send button with nobody checking whether it was even accurate, that’s the actual problem, and it would be the actual problem even without AI involved.
I want to name the distinction I think people are actually reacting to, because “AI-generated” gets used as shorthand for two very different things. One is content a person shaped, decided on, and stands behind, that happened to be drafted with a tool. The other is content nobody really authored, that got generated and shipped with no human genuinely in the loop. Readers can feel the difference even when they can’t name it. The first feels like a person used a better pen. The second feels like nobody was home. When someone says they don’t trust AI content, they’re almost never objecting to the pen. They’re objecting to the sense that nobody was home, and yes, they’re right to.
The trust data on this is worth sitting with. A 2026 Klaviyo survey of 8,000 consumers across eight countries found that when people notice AI-generated content from a brand, they’re roughly four times more likely to trust that brand less than more, 31% reported decreased trust versus 7% who said it increased, according to The State of Brand. The same report found half of consumers can now correctly spot AI-generated content on sight. People aren’t fooled nearly as often as businesses seem to hope, which means the move isn’t hiding the tool. It’s making sure there’s a real person behind the thing whether anyone can tell or not.
This is also exactly why I keep coming back to sourcing and hallucination in these issues. An AI tool that states something as fact without a real source behind it isn’t a shortcut, it’s a liability wearing your name. Every stat I’ve shared in this newsletter has a real link behind it, because the moment I stop being able to say where a number came from, I’ve crossed from helpful into just sounding confident, and confident and true have never been the same thing, no matter how good the sentence sounds.
Takeaway: Nobody minds that AI helped. They mind finding out nobody actually checked.
Try this: Look at the last thing you sent that AI helped draft. Ask yourself honestly, did a real person actually check it was true and actually you before it went out, or did it go straight through? If it went straight through, that’s not a reason to feel bad, it’s a reason to add one step: a real read, by a real person, before send. That single step is the whole difference this issue is about.
Ask me anything: This month’s open question, since I’d genuinely rather hear from you than just talk at you: what’s the one part of your marketing you’re most nervous to hand to AI, and why? Reply and tell me, I read every one, and it might become a future issue.
Thanks for being here. Talk soon, Tiffany