Influencer Marketing Has an AI Fake Problem, and Brands Are Pretending Not to Notice.

I sat in a pitch meeting recently where an agency proudly presented an influencer’s engagement rate as a headline metric, no context, no breakdown, just a big confident number on a slide. I asked how much of that engagement had been checked for bot activity or AI-driven comment farming. The room went quiet in a way that told me everything.

That silence is the influencer marketing industry right now. Everyone can see the problem. Almost nobody wants to be the one who says it out loud, because a lot of budget is currently flowing based on numbers that don’t hold up to scrutiny.

Two Separate Problems, Same Root Cause

Fake engagement and fake influencers, both getting harder to spot

First, there’s engagement fraud, which isn’t new but has gotten dramatically more convincing. AI-generated comments that sound plausibly human, engagement pods running on increasingly sophisticated automation, follower counts padded with bot accounts that behave realistically enough to survive a casual audit. The tools that used to catch obvious fake engagement were built for obviously fake engagement. This isn’t obvious anymore.

Second, and this one’s newer, fully virtual influencers with AI-generated faces, voices, and personalities, some disclosed, plenty not clearly disclosed, building audiences that brands are paying real money to reach. There’s nothing inherently wrong with a virtual influencer done transparently. There’s a real problem when brands, and often the audience itself, don’t actually know what they’re engaging with.

Why Brands Keep Looking Away

Nobody wants to audit a number that’s making their report look good

Here’s the incentive problem, plainly. If a campaign report shows a great engagement rate and a low CPM, questioning that number risks turning a good-looking result into a bad-looking one. Marketing teams under pressure to show results have a quiet incentive not to dig too hard into whether the underlying number is real. That’s not malice, it’s just how incentives work when nobody’s specifically rewarded for finding uncomfortable truths in their own reporting.

What I’d Actually Do

Demand raw data, or walk

• Ask for raw analytics access, not agency-summarized screenshots, before committing to any influencer partnership at meaningful spend.
• Run independent audience quality checks using third-party tools, don’t rely solely on the influencer’s own platform-reported numbers.
• Require clear disclosure on whether an account is fully or partially AI-generated. If a partner won’t answer directly, that’s your answer.
• Weight conversion and business outcomes far more heavily than engagement rate in how you evaluate influencer performance going forward. Engagement is the easiest number to fake. Revenue is much harder to.

Quick Answers

A few direct questions, answered directly

How common is fake engagement in influencer marketing?
It remains widespread, and AI has made fraudulent engagement significantly harder to detect through casual review, since bot-driven comments and interactions increasingly mimic genuine human behavior.

What are virtual influencers?
AI-generated personas, including synthetic faces, voices, or personalities, used to build audiences and partner with brands. Some are clearly disclosed as virtual, others are not, which raises transparency concerns.

How can brands verify influencer authenticity?
By requesting raw platform analytics rather than summarized reports, using independent third-party audience auditing tools, and prioritizing verified conversion outcomes over engagement metrics alone.