Authenticity is hard currency now. Everything else is inflation.

Content used to be scarce, so producing it was the hard part. Now content is nearly free, and a machine can generate more of it in a minute than a team once produced in a quarter. What's scarce again is something else entirely: a real person, describing a real problem, in their own words. That's what's actually worth paying attention to.

AI-generated writing now accounts for roughly half of new English-language articles online – up from under 8% when ChatGPT launched in late 2022 – according to research firm Graphite. Every channel carries some share of text nobody wrote, in the sense that used to matter.

The audience noticed. In a 2026 survey from Sprout Social, more than half of respondents said they see so-called "AI slop" often or very often in their feeds, and 88% said their trust in social media news has fallen as a direct result. Video tells the same story: in Animoto's 2026 State of Video Report, 83% of consumers said they can identify AI-generated video on sight, and more than a third said spotting it lowers their trust in the brand behind it. Fluency was never the same thing as credibility. Once fluency became free, the two started pulling apart.

Three tells.

People didn't get better at reading minds. They got better at reading absence – the specific things missing from content nobody actually lived through.

1. Specificity. A real account has an odd, unnecessary detail – the exact objection a customer raised, the number that surprised everyone in the room. Generated content smooths those details out, because it has none to draw on.

   
   

2. Friction. A real perspective includes a trade-off, a wrong turn, something that didn't work. Generated content resolves too cleanly, because it isn't reporting from anywhere.

   
   

3. Stakes. A real person loses something by being wrong – credibility, a client relationship, standing among peers. Polished brand copy risks nothing and readers have learned to price that in.

   
   

None of this is abstract for B2B specifically. It shows up directly in buyer behavior.

The premium is measurable.

In a July 2025 study of B2B buyers across 14 markets, commissioned by LinkedIn, 77% of marketing leaders said their buyers still turn to personal networks – not brand channels, not AI tools – to vet a vendor. Among the youngest buyers, the number holds exactly as high: 77% of 18-to-24-year-olds say no amount of AI replaces the judgment of someone they trust.

Employee-shared content follows the same pattern: research on B2B advocacy programs consistently finds that a message shared by an employee earns roughly three times the credibility of the identical message run as a paid ad – because a name and a face are attached to it, and a colleague, not a company, is the one answering for it.

Scarcity, not sentiment.

None of this is really about warmth. It's about supply. For twenty years, marketers optimized for polish because polish was expensive and scarce, and a slick campaign signaled real investment behind it. AI broke that signal. Polish is now the cheapest thing in the building – any competitor can generate it by lunchtime. What's expensive is proof that a real person was actually there: that someone lived the problem, said the specific thing, and is willing to put their name behind it.

That's the trade the market just made, whether anyone announced it or not. Content got cheap. Verified human experience got expensive. Keep spending on the first while a competitor spends on the second, and you're funding a currency that's losing value every week.