Real or AI: Why Audiences Are Fleeing User-Generated Content for Trusted Sources

Open any feed today and the same question keeps echoing in the back of your mind:

Is this real, or AI?

That TikTok “review,” the earnest YouTube testimonial, the five-star Amazon comment written in flawless corporate speak – they all blur into the same slightly uncanny tone of generative AI. What used to be the raw, human layer of the internet – user-generated content (UGC) – is now under suspicion.

The result is a quiet but measurable shift: audiences are losing faith in random posts from strangers and drifting back toward recognizable brands, newsrooms, and entertainment networks that can stake their reputation on getting things right.

This isn’t just a mood. The numbers back it up.

The trust crash: when the feed stops feeling human

A 2025 survey of 2,000 U.S. adults found that people now believe only 41% of what they read online is both accurate and human-generated. The rest is seen as partially unreliable (36%) or outright false (23%). Three-quarters of respondents said they trust the internet less than ever, and 78% say it’s getting harder to tell real content from AI-generated posts.

In the UK, a national attitudes study found 72% of people are unsure online content can be trusted because it may be AI-generated, and 78% are worried about the negative outcomes of AI.

People are trying to adapt – scanning, second-guessing, hovering over every “authentic” story with skepticism – but they don’t feel in control. That’s a terrible foundation for UGC, which historically relied on an unspoken assumption: if it looks like a real person talking, it probably is.

How AI broke the old UGC deal

For about 15 years, UGC was the antidote to polished corporate messaging:

YouTube reviews beat TV ads.
Reddit threads beat press releases.
TikTok “day in the life” posts beat brand campaigns.

The social contract was simple: brands spin, people tell the truth.

Generative AI has shredded that simplicity. The same tone and formats that made UGC powerful are now trivially easy to simulate at scale. A small brand—or a bad actor—can generate thousands of synthetic “personal experiences” overnight: fake reviews, scripted comment threads, influencer-style shorts, even simulated arguments for engagement.

The Stanford 2025 AI Index notes that global confidence in AI companies doing the right thing with people’s data is actually falling, dropping from 50% in 2023 to 47% in 2024, along with declining belief that AI systems are fair and unbiased.

So we’ve landed in a paradox: AI makes it easier than ever to impersonate “real people” online, just as trust in the companies behind that AI is sliding. UGC hasn’t disappeared, but the baseline assumption that it’s human, unbiased and authentic absolutely has.

Entertainment’s warning shot: music you can’t tell apart

The entertainment world is already living in the next phase of this problem.

In late 2025, Deezer and Ipsos ran a survey across eight countries and 97% of listeners could no longer reliably distinguish AI-generated music from human-made tracks in blind tests. Over half (52%) said this made them uneasy. At the same time, 80% want clear labeling for AI-generated songs, 45% want the ability to filter them out, and 52% oppose AI-only tracks being eligible for official charts.

Deezer now says around 34% of tracks delivered to its service each day are fully AI-generated, or about 50,000 songs daily.

Listeners aren’t rejecting AI outright; they’re curious. But they’re drawing a line when it comes to trust:

They want to know what’s real and what isn’t.
They don’t want AI tracks silently competing with humans for chart positions and royalties.
They expect platforms to step in with labelling and filtering.

That’s exactly the psychology now spreading to text, video and “authentic” UGC.

From “trust the internet” to “trust specific brands”

At the macro level, trust is fragmenting rather than disappearing.

The 2024 Edelman Trust Barometer shows that business is currently the only global institution that sits firmly in the “trusted” zone, with around 63% of respondents saying they trust business to do what’s right. That’s “far ahead of government” on both competence and ethics in Edelman’s wording.

It doesn’t mean audiences suddenly love corporations; it means they see named entities with something to lose as more reliable than the amorphous “internet.”

News is following a similar pattern. The Reuters Institute’s Digital News Report 2024 finds that overall trust in news remains fragile (around 40% globally say they trust “most news most of the time”), but news brands still outrank social media and “people like me” as sources for verifying information. When users see something suspicious online, the most common behaviour is to check it against a news outlet they know.

Meanwhile, a YouGov study on generative AI in media found that 7 in 10 Brits trust news content generated by AI less than news written by humans. Only 7% trust AI-generated news more.

Put that together and a new pattern emerges:

People trust “random content on the internet” less.
They still trust named news brands more than the feed.
They trust business and their own employer more than they trust anonymous peers and unknown creators.

The more AI muddies the baseline of UGC, the more valuable these named, accountable entities become as anchors of reality.

Platforms’ existential problem: polluted UGC

For social platforms whose whole pitch is “real people talking to real people,” this is an existential issue.

The Talker Research / World survey paints a stark picture:

Americans suspect AI is behind almost half of social media posts they see.
Only 31% feel confident telling AI-generated reviews from human ones.
When tested, just 30% could correctly identify whether business reviews were written by humans or AI.

When the average user assumes that half of what they see in their feed might be synthetic, anonymous “UGC” stops feeling like a shortcut to the truth and starts feeling like an attack surface.

At that point, UGC splits into two very different categories:

Open UGC: low-friction, low-verification content that anyone (including bots) can publish. High volume, low trust.
Curated, verified UGC: content from identifiable creators, real communities, or brand-backed environments where identities and incentives are clear. Lower volume, higher trust.

The open end of the spectrum will always exist. But the more AI floods it, the more users and advertisers will migrate to the curated side.

Brands and networks as “trust routers”

For tech, media, and entertainment companies, this shift is an opportunity as much as a risk.

If you’re a news network, streaming platform, sports league, game publisher, or major entertainment franchise, you’re no longer just competing on IP and UX. You’re competing on:

Signal vs noise: how much low-quality synthetic content you allow into your ecosystem.
Transparency: whether you clearly label AI-assisted or AI-generated content.
Editorial courage: what you exclude – from charts, front pages, recommendation carousels.

We’re already seeing early moves:

Music platforms like Deezer tag AI-generated tracks and keep them out of editorial playlists to protect human artists and maintain listener trust.
Publishers and broadcasters experiment with “human-written” labels, provenance metadata and stricter AI house rules for news and opinion pieces.
Some brands introduce “verified reviewer” programs and identity-based reputation to separate real customers from synthetic review spam.

In effect, these entities become trust routers. Audiences rely on them not just to entertain and inform, but to:

Filter out the worst of the synthetic sludge.
Vouch that the people behind the content actually exist.
Tell users when AI is in the loop – and how.

The brands that lean into this, instead of quietly auto-generating more content to “keep up,” will be the ones audiences keep returning to.

What this means for creators

Independent creators are caught in the crossfire. They’re competing not only with each other, but with AI models that work 24/7 and never need a day off.

That doesn’t mean creators are doomed. It means the old playbook of faceless, generic “UGC-style” output is losing value fast.

The creators who will win in a Real or AI world are the ones who:

Show proof of life: behind-the-scenes footage, live streams, in-person events, and time-stamped, verifiable work.
Build persistent identities: consistent voice, long-term presence, and a body of work that would be very hard to fabricate overnight.
Partner with trusted brands and networks: co-productions, official channels, and curated communities that lend institutional trust to personal voices.

In other words, they stop relying on the default credibility of “I’m just a person online” and start building explicit trust signals on top.

The next era: smaller, sharper, more branded

The story of the past 15 years was scale: more creators, more platforms, more UGC, more everything.

The story of the next decade might be contraction: not fewer creators, but fewer places audiences truly trust.

As feeds fill with AI-generated noise, audiences will still scroll. But when it matters—news, money, health, voting, even which show to watch next—they’ll increasingly look for:

A channel bug or studio logo.
A trusted news brand.
A verified creator they’ve followed for years.
A platform that visibly treats AI as a tool to manage content, not as a firehose to flood users with it.

UGC isn’t going away. But the default assumption that “user-generated” equals “real” is over.

In its place, we’re moving into a world where trust is a premium feature. And the biggest winners of the Real-or-AI era will be the brands, newsrooms, and entertainment networks that treat authenticity not as an aesthetic, but as infrastructure.