UGC

AI UGC Is Losing: Preference Fell From 60% to 26%

By Sean HolleranAugust 27, 20268 min read
Humanoid robot seated alone on a bench

For about two years the pitch has been the same: AI creators are coming for UGC. Why pay a person to film your product when a model can generate an infinitely patient spokesperson who never asks for usage rights?

The data went the other way. Not slightly — dramatically, and fast.

Consumer Preference Collapsed

In research from Billion Dollar Boy, 60% of consumers in 2023 said they preferred generative-AI creator content over content from traditional creators. By the time Digiday reported the follow-up wave in January 2026, that figure was 26%.

60%
Preferred AI creator content, 2023
26%
Preferred AI creator content, Jan 2026
25%
Of marketers use AI in influencer work at all

That is a 34-point drop in roughly two to three years. It is one of the sharper reversals in consumer sentiment toward a technology that anyone has measured, and it happened while the underlying technology was getting better, not worse.

Which tells you the objection was never quality.

Saturation Did It

The mechanism is not mysterious. When AI content was rare, it read as novel. When it became the default texture of the feed, it started reading as cheap — and then as evasive. The Guardian found more than a fifth of the videos shown to new YouTube users are now AI-generated filler. Consumers learned the visual signature, and once you can spot it, you cannot unspot it.

The reaction inside the influencer industry has been striking. Brands are now writing "no AI at any stage" into creator briefs. And they are actively requesting the things production used to sand off — unmade beds, wrinkled clothes, a messy background, a take that is not quite clean.

Imperfection stopped being a flaw and became the proof of authorship.

Becky Owen, CMO at Billion Dollar Boy, described the shift as a reset, saying consumer sentiment against AI is "roaring." Whatever you make of the phrasing, the direction is consistent across every dataset we could find.

Marketers Never Actually Adopted It

The other half of the story is that the industry-wide AI-UGC adoption wave mostly did not happen.

A Digiday survey of 100+ marketing professionals fielded in Q1 2026 found only 25% use AI for influencer marketing work at all — and of those who do, most use it for data analysis and outreach rather than making the content. A WFA survey of brand marketers found 60% have no plans to use AI influencers at all, and 96% cite concerns about consumer trust and acceptance — though on a small sample of 33 respondents across 27 brands.

So the picture is not "brands love AI creators, consumers hate them." It is that both sides looked at it and mostly declined.

And Now Meta Puts a Label On It

This is the part that turns a sentiment problem into a structural one.

Under Meta's current transparency policy, an "AI info" label is applied automatically for most commercial ads — advertisers do not self-declare, though political and social-issue advertisers still must, and Meta notes not every ad touched by its AI features gets a label. Usually it sits behind the three-dot menu. But when an ad contains an AI-generated photorealistic human, the label appears in-feed, right next to "Sponsored."

That is exactly the case for an AI spokesperson ad. And under this year's policy update, Meta also automatically detects ads made with third-party AI tools through industry-standard provenance signals, and labels those too.

The Practical Consequence

You do not get to decide whether the viewer knows. If your ad features a synthetic person, the platform is increasingly telling them so in the feed — and roughly three-quarters of consumers no longer prefer that content. You are volunteering into a disclosed disadvantage.

Worth noting what is not true, since it is circulating: there is no confirmed Meta rule requiring commercial advertisers to disclose AI-generated content. The requirement we could verify covers political and social-issue advertising. The labeling above is automatic detection, not a compliance obligation you can satisfy with a checkbox.

Where This Leaves the Honest Case for AI

We use AI tools daily. This is not a purity argument, and pretending AI has no role in ad production would be its own kind of dishonesty. The distinction that holds up in the data is who is on camera.

What works:

What does not:

The Format Data Backs It Up

Motion's 2026 creative benchmarks, drawn from $1.29 billion in Meta spend across 578,750 creatives, ranked UGC fourth out of every asset type on hit rate at 7.56% — behind text-only at 11.60%, product image with text at 8.75% and lifestyle product imagery, and ahead of high-production video at 6.97%.

We will be straight about that, because it complicates our own pitch: UGC is not the highest-odds format per asset, and it is not second or third either. What it is, is a format that consistently beats expensive polish while costing a fraction of it, and that can be produced at the volume Meta's auction now demands. High production sits mid-pack on hit rate and near the bottom on spend efficiency. That is the trade that matters.

The through-line across all of it — text-only winning, high production underperforming, AI spokespeople getting labeled and distrusted — is that Meta's feed rewards things that look like they came from a person with something to say, and punishes things that look manufactured. Expensive manufactured and cheap manufactured both lose.

The Bottom Line

The AI-UGC thesis assumed consumers would not notice or would not care. They noticed, and by January 2026 preference had fallen from 60% to 26%. Meanwhile the platform started labeling synthetic humans in-feed, and 96% of brands avoiding virtual influencers say trust is why.

Use AI for everything except the face. The face is the entire asset.

If you want ads with real creators in them, produced at the volume the auction actually requires, that is what we do — or see the work first.


Sources: Digiday, "After an oversaturation of AI-generated content, creators' authenticity and messiness are in high demand" (Jan 14, 2026); Digiday+ Research, "Marketers hesitate to adopt AI for influencer and CTV marketing" (Jul 9, 2026), surveying 100+ marketing professionals in Q1 2026, citing WFA (Apr 2025); Billion Dollar Boy consumer research via Digiday; The Guardian on AI video volume; Meta generative-AI ads transparency page (2026 update); Motion Creative Benchmarks 2026 ($1.29B Meta spend, 578,750 creatives, Sept 2025–Jan 2026 window).

Real Creators. Real Numbers.

UGC video ads with actual people in them, produced at the volume Meta’s auction demands.