Your Customers Can Spot AI Content — And They’re Punishing Brands That Use It Badly

You think you can spot AI-generated content. You probably can't.

When researchers showed 301 US adults a mix of real photographs and AI-generated images in September 2025, participants were correct about 50% of the time — no better than a coin flip. And it's getting worse. The same study ran in 2023 and 2024. Detection accuracy has dropped as generation quality has improved. Consumer confidence, meanwhile, has gone up.

And yet: consumers insist they notice AI content. 82% say they spot AI-generated writing often or always. In the 22–34 cohort, that jumps to 88%.

Both things are true. People feel the AI. They often can't prove which specific piece is AI. And when they feel it — deserved or not — they punish the brand anyway.

The Trust Penalty Is Real and Measurable

Across four 2024–2026 studies, the pattern converges: detection is fuzzy; penalty is sharp.

The 60% number is the one to sit with. It's not "lose interest in a post" — it's "close the browser tab and don't complete the transaction." That's revenue that was real yesterday and isn't today, because 88% of Americans now say it's harder than a year ago to tell what's real online.

When consumers do detect AI copy in social media, they describe the brand in specific terms — 20% call it untrustworthy, 20% call it lazy, 19% call it uncreative. This isn't reputation damage in the abstract. This is your customers looking at what you shipped and deciding you didn't care enough to show up.

Detection is fuzzy. Penalty is sharp. The business risk doesn't wait for the audience to be right.

What AI Slop Actually Looks Like

You've seen it. You might be publishing it. Here's how to tell.

AI slop isn't bad grammar or obvious robot speak. That era is over. Modern AI writes fluently. The problem is what's missing — specificity, local knowledge, actual opinions, the kind of detail that only comes from someone who was there.

The test is simple: could this content have been written about any business in your industry, in any city, without changing a word? If yes, it's slop. It doesn't matter how polished the prose is. Polish without substance is exactly what consumers are learning to detect.

The Law Is Catching Up

Consumer instinct is one thing. Regulation is another. Both are pointing in the same direction.

California's AI Transparency Act (SB 942) was originally set to take effect January 1, 2026. In October 2025, Governor Newsom signed AB 853, pushing the operative date to August 2, 2026 and aligning the timeline with the EU AI Act.

When it takes effect, SB 942 will require AI providers to embed "latent disclosures" — invisible digital markers — in AI-generated content. The markers identify the provider, timestamp, and system that created the content. Penalties run up to $5,000 per violation per day.

The FTC is already enforcing in an adjacent space. Its rule banning AI-generated fake reviews and testimonials took effect October 21, 2024. Violations carry civil penalties of up to $51,744 per violation, adjusted annually. The FTC's Operation AI Comply enforcement track has already brought actions against companies for undisclosed AI use.

Detection infrastructure is also maturing fast. Deloitte projects the deepfake detection market will grow 42% annually, from $5.5 billion in 2023 to $15.7 billion in 2026. OpenAI's own DALL-E 3 classifier correctly identifies around 98% of DALL-E 3 images in internal testing — though it only detects that one generator, not the broader category. The question isn't whether your audience will know. It's whether you'll have gotten ahead of it.

Transparency Is the New Advantage

Disclosure doesn't hurt engagement. It measurably improves it.

When Yahoo and Publicis Media tested AI-generated ads with visible "AI-generated" disclosures against the same ads without them, the disclosed ads outperformed on every trust metric measured.

The underlying signal picked up by Getty Images' VisualGPS research across 25 countries explains why disclosure works: 98% of consumers say authentic images and video are pivotal to establishing trust with a brand. 76% say they've reached the point where they can't tell if an image is real anymore.

Put those two findings together. Authenticity matters more than ever. Detection is harder than ever. The brand that says "we used AI to draft this and our team verified every claim" resolves the ambiguity for the audience — and the data says they reward it.

Think of it as the evolution of authenticity in marketing. Each era raised the bar on what "real" means — and each time, the businesses that moved first captured the trust premium.

Authenticity 3.0 isn't about avoiding AI. That ship has sailed. It's about being transparent about how you use it. Tell your audience what AI does in your process and what humans do. Show where editorial judgment lives. The companies that make that visible first define what responsible AI use looks like in their market. Everyone else spends time explaining why they didn't.

The businesses that disclose their AI use first will own the trust premium. Everyone else will be explaining why they didn't.

How We Handle It at ShipsMind

We use AI as infrastructure. We don't let it replace judgment.

Every piece of content we produce — including the post you're reading right now — goes through the same pipeline: AI handles research synthesis, first drafts, and data gathering. Human editorial judgment handles voice, accuracy, local specificity, and the decision about whether something is worth saying at all.

We don't publish what a model produces. We publish what we decide to say, using models to help us say it faster and with better-researched backing. The difference matters. It's the difference between a contractor who uses power tools and a power tool that builds houses by itself.

Three Things You Can Do This Week

You don't need to overhaul your content strategy overnight. Start here.

Frequently Asked Questions

Common questions about AI content, trust, and what to do about it.