AI Content vs Human Content: What the Data Says About Trust
A study by Bynder, 2,000 people in the UK and US were shown 2 articles on the same topic. One was written by ChatGPT, the other by a professional copywriter. Unlabelled, 56% said they preferred the AI version.
But when people suspected content was AI-made, 52% said they'd become less engaged with it
That's the whole AI content debate in 2 numbers. The writing quality holds up fine. The trust hit happens when people spot it.
How many brands are using AI to make content?
Most of them. HubSpot's 2026 State of Marketing report found 42.5% of marketers use AI extensively for content creation and another 38% use it occasionally. Validity's research, covered by MarTech, found 74% of marketers are deploying or testing AI-generated content, and 43% plan to spend more on it this year.
When most of your competitors lean on the same tools, feeds start to look the same. Same hooks, same structure, same "In today's world" openers.
Can people actually spot AI content?
About half can. In the Bynder study, 50% correctly identified the AI-written copy. US participants did better (55%) than UK participants (45%).
Once they spot it, the brand takes the hit. Asked how they'd feel about a brand posting AI content on social:
Reaction | % of consumers |
Brand feels impersonal | 25% |
Brand feels untrustworthy | 20% |
Brand feels lazy | 20% |
Disclosure is now a trust issue
Sprout Social surveyed more than 2,000 consumers in the UK, US and Australia in Q3 2025. 52% named brands posting AI-generated content without disclosure as their top concern on social, tied with brands mishandling personal data.
Email tells the same story. 40% of consumers told Validity they'd trust a retailer's marketing emails less if they knew AI wrote them. Only 25% said it would increase their trust.
What people are fine with
The good news for anyone using AI sensibly: 82% of consumers agreed they don't mind brands using AI to help write copy, as long as it feels like a person wrote it.
So the line people draw is simple. AI helping behind the scenes is fine. AI replacing the human voice they followed you for is where you lose them.
How we use AI without losing trust
This is how we approach it for clients:
Strategy is human. A person decides what the brand is saying this month and why.
AI drafts. We use it for first drafts, variations and repurposing, working from each brand's saved voice guidelines.
A person rewrites. Every post gets edited for voice, local references and the real detail only someone close to the business would know.
Facts get checked. Any stat, date or claim is verified against the original source.
Replies stay human. Comments and DMs are where trust is won or lost, so people handle them.
The test we use: if a loyal customer read this post, would they believe it came from you? If not, it goes back.
Should you label AI content?
If an image or video is fully AI-generated, label it. Most major platforms now have AI labels or disclosure settings, and the Sprout data shows undisclosed AI is what bothers people most. For text that AI helped draft and a person rewrote, most brands don't label it, the same way nobody labels a spell-checker.
FAQs
Is AI-generated content bad for engagement? Only when people notice it. 56% preferred an unlabelled AI article, but 52% disengage when they suspect AI.
Do consumers trust AI content? Mostly not when it's obvious. 25% find it impersonal and 20% untrustworthy on social. 82% are fine with AI helping, as long as the result feels human.
Should brands disclose AI use on social media? For fully AI-generated images and video, yes. It's the top consumer concern in Sprout Social's Q3 2025 data.
How we can help
Not sure if your feed reads as human? 2 routes:
Done for you: request a free social media audit and we'll flag any posts that read as AI, plus what to change.
Learn it yourself: The Modern Business Method COURSE teaches you, or your whole team, to use AI without losing your voice: https://tartansocial.mykajabi.com/the-modern-business-method-sales-page
Next in the series: how to use ChatGPT for social media, with prompts you can copy.
Comments