That 'AI-Generated' Label on Your Ad? It's Hiding More Than It's Telling You


That 'AI-Generated' Label on Your Ad? It's Hiding More Than It's Telling You

 

You see it now on half the ads scrolling past you. A little tag: "AI-generated."

What does that actually tell you? Less than you'd think.

I've spent years figuring out how things work before I trust them — whether that's a sketchy border crossing in Central America or a new piece of tech promising to make my life easier. So when I saw researchers digging into what "AI-generated" labels really mean, I got curious. Turns out, that label is doing a lot less work than most of us assume.

Here's the problem. AI could mean a designer used it to tweak a background or clean up a photo — the kind of thing a human editor has done with digital tools for decades. Or it could mean AI built almost the entire ad from scratch, with a human barely touching it. The label "AI-generated" covers both situations. And everything in between. Same two words, wildly different stories.

Why does that matter to you? Because trust is on the line. Research shows that when companies disclose AI involvement, it can actually lower how much people trust the ad. Makes sense — we're wired to be a little suspicious of ads anyway. Add "a machine made this" and suspicion ticks up. But reactions aren't uniform. How much people push back depends heavily on what the AI was used to create.

Here's where it gets interesting. We've had disclosure in advertising forever — think "#ad" on an influencer's post, or "sponsored content" tags. Those labels tell you why you're seeing something: because someone paid for it. AI labels do something completely different. They tell you how something was made, not why you're seeing it. And they don't break down the split between human and machine. You're left guessing.

That gap in information helps explain why brands like Coca-Cola and Gucci caught heat over AI-led campaigns. People weren't just reacting to "AI was used." They were reacting to not knowing how much creative control a human actually kept. Did someone develop the concept and just let AI polish it? Or did AI dream up the whole thing while a human clicked "approve"? Nobody could tell from the label alone.

Governments are starting to notice this gap. The EU is pushing stricter disclosure rules for certain generative AI systems. There's also new technology in the works that would let digital content — ads included — carry a traceable history of how it was edited and produced. That's a step toward actual transparency instead of a vague sticker.

Meta already learned this lesson the hard way. Its "Made with AI" label was getting slapped on content that had only minor AI tweaks, without giving people enough context to understand what that meant. Meta eventually swapped the wording to "AI info" and dialed back how prominently the label showed up when AI was just editing, not generating, the content. Small wording change. Big difference in how people interpreted it.

And it's not only the AI labels causing confusion. Now we're seeing "human-made" and "no AI" tags pop up too — meant to reassure people. But do those mean zero AI involvement anywhere in the process? Or just that a human led the creative work with AI quietly assisting somewhere behind the scenes? Same ambiguity, flipped around.

So what do you do with all this the next time you spot an "AI-generated" tag? Don't treat it as a verdict. Treat it as a question. It tells you AI played some role — not whether the content is good, bad, fake, or trustworthy. It doesn't tell you who had the idea, who refined it, or who checked the facts before it went live.

I've learned the same lesson applies well beyond advertising — with new tools, new countries, new situations. A label or a headline gives you a starting point, not the whole picture. The real answers come from asking "how exactly did this happen?" instead of settling for the short version.

Next time you see that little AI tag, don't just nod and scroll. Ask the follow-up question. That's where the real story usually is.

Source: The Conversation

See also our earlier piece: Why ChatGPT Might Be Ignoring Your Website (And What One New Study Says To Do About It)