A man in orange robes sits cross-legged before a Buddhist altar, speaking softly about Qi, inner harmony, and balance. On Instagram, 2.6 million people follow him. In 90 days, he generated over $300,000 selling ebooks and digital wellness classes. His name is Yang Mun. He doesn't exist.

Every frame of Yang Mun is synthetic — face, voice, robes, and altar — built by a solo creator in Israel chaining ChatGPT, ElevenLabs, HeyGen, and Google's video models (confirmed by Google's SynthID watermark detector). Millions of people watched, trusted the robe, and paid money to learn from a fiction.

TL;DR

Synthetic influencers are already scaling into six-figure businesses because engagement algorithms route AI video straight to the older demographics least equipped to spot it.

  • The Yang Mun blueprint — One solo operator generated 400M+ organic views, 2.6M Instagram followers, and $300K in profit in 90 days with a 100% synthetic Buddhist monk.
  • The demographic asymmetry — Across 6.3M views on our own tabiji.ai Instagram reels, 76.4% of viewers are 45+ (and under 1% are under 25).
  • Zero platform incentive to label — Models like Veo 3.1 and Seedance 2.0 have erased visual tells, and platforms won't enforce visible AI labels because synthetic video drives watch time.

The Monk Who Wasn't Real

Yang Mun hit 2.5 million Instagram followers in three months, added hundreds of thousands more on TikTok, and crossed 400 million organic views across platforms.

Yang Mun's Instagram profile showing 2.5 million followers and AI-generated monk content
Yang Mun's Instagram profile — 2.5 million followers, 268 posts, all AI-generated.

The formula was surgical: pick an archetype that projects calm authority (an elderly monk in saffron robes), script bite-sized aphorisms on wellness and traditional Chinese medicine, and let saves and shares compound in the Reels algorithm.

Then monetize. Yang Mun's operator charged $50/month for wellness classes plus digital guides, netting $300K in 90 days with zero staff and near-zero marginal cost.

When the creator, Shalev Hani, eventually stepped forward, he billed himself as a “Digital Creator & AI Storyteller.” In his framing, Yang Mun wasn't a con — it was a character, no different from an animated series. That blur between entertainment character and trusted advisor is exactly what makes synthetic influencers so lucrative and so dangerous.

Who Is Actually Fooled

I run tabiji.ai's Instagram account, where we publish AI-generated travel reels covering destination tips and cultural etiquette. Over a recent 30-day window, our reels pulled 6.3 million views and 1,800 new followers, with 98.8% of reach coming from non-followers via algorithmic recommendations.

Tabiji Instagram insights showing 6.3 million views in 30 days Tabiji audience demographics showing 76% of audience is 45 and older

Look at the age breakdown on those 6.3 million views: 76.4% of our viewers are 45 or older. The 55–64 bracket alone is 30.2%, and 65+ is 23.3%. Viewers under 25 — the cohort that grew up with face filters, deepfakes, and generative tools — make up less than 1%.

Recommendation algorithms are systematically funneling synthetic video to the oldest cohorts on the internet. Not out of malice, but because older users watch longer, save more, and leave earnest comments — and the feed ranks on engagement.

Users under 30 built a default skepticism from a decade of filtered influencers and bot accounts. Users over 55 spent most of their lives in a media environment where video footage meant a camera pointed at a real person in a real room. Recommendation engines are monetizing that assumption.

The Tools Are Too Good Now

Eighteen months ago, you could spot synthetic video by the six-fingered hands, plastic skin texture, or robotic cadence. That window is closed.

Google's Veo 3.1 renders 4K vertical video with native audio and continuous scene extensions past 60 seconds; every Google account gets 10 free generations a month. ByteDance's Seedance 2.0 handles phoneme-level lip sync across 8+ languages with multi-shot narrative control from a single prompt, wired directly into CapCut.

Saying “I can always tell when it's AI” is a 2024 cope. At scroll speed on a phone screen, most people can't — especially not a 60-year-old scrolling Reels at 11 PM.

The Commercialization of Fake People

Synthetic personas aren't just a solo-hustler trick; they are already industrialized ad tech.

MakeUGC content library showing AI-generated people in UGC ad formats
MakeUGC's content library — AI-generated "creators" ready to pitch your product in any niche.

Platforms like MakeUGC ship catalog libraries of synthetic humans staged in realistic bedrooms, kitchens, and cars. Pick an avatar, paste a script, and render a “customer testimonial” for skincare, supplements, or SaaS in two minutes. For DTC brands, the unit math is obvious: why pay ten human UGC creators $300 a clip when you can A/B test fifty synthetic hooks for $20?

And the exact same stack that lets a DTC brand test ad hooks lets an anonymous operator spin up a synthetic holistic doctor, a fake fitness transformation, or a fictional wealth manager.

What's Coming Is Worse

Selling $50 breathwork ebooks via a synthetic monk is the mild version. Here is what the next 12 months look like:

90-day synthetic body transformations. An operator generates a slightly overweight avatar on Day 1, posts daily AI-rendered gym clips and weigh-ins for three months, and reveals a shredded physique on Day 90 — followed by a link to a $99/month supplement subscription. The transformation is 100% pixels; the credit card charges in the comments are real.

Synthetic financial and legal advisors. Articulate, silver-haired avatars in wood-paneled offices posting daily market breakdowns, funneling retirees into WhatsApp trading groups or high-fee annuities.

Synthetic local authority accounts. City-specific “local foodies” and neighborhood guides run entirely via API, monetizing local lead-gen and restaurant placements without ever setting foot in the city.

Platforms Won't Fix This

Meta, TikTok, and YouTube can detect synthetic uploads easily — between C2PA metadata, SynthID watermarks, and acoustic fingerprinting, the signals are sitting right in their ingestion pipelines.

They don't throttle it because synthetic content juices engagement metrics.

AI accounts publish higher-contrast visuals at 10× the cadence of human creators, keeping watch time high and ad inventory full. Slapping a prominent, unskippable “AI-Generated Persona” banner across every synthetic Reel would cause viewers to swipe past — directly cutting session time and quarterly ad revenue.

Platforms will not enforce visible, friction-inducing disclosure until regulators force their hand. The EU AI Act will bite first; US enforcement will lag by years.

The New Affiliate Marketing

I'm not writing this from the sidelines. I run automated content pipelines, test MakeUGC and WaveSpeed daily, and wrote The Future Is Synthetic. Operating inside the kitchen is what makes the trajectory obvious.

As AI Overviews hollow out traditional blog SEO, affiliate marketing is migrating into short-form video. Video has ten times more levers than text — hook framing, voice cadence, phoneme lip-sync, B-roll pacing, background score — creating a massive arbitrage window for operators who master Veo 3.1 and Seedance 2.0 early. To Gen Z operators, programmatic AI video in 2026 is what niche-site SEO was in 2015.

The operators who disclose what they're building will build durable media brands. The ones running unlabeled medical, financial, and spiritual cons on 65-year-olds will trigger the regulatory backlash that brings the hammer down on the whole ecosystem.

Fake Until Proven Real

For thirty years, the internet's default heuristic was real until proven fake. A video of a person talking to a camera was evidence that a human sat in front of a lens.

That heuristic is dead. The only rational posture left is fake until proven real — and our cryptographic provenance infrastructure is years behind the generation tools:

  • Mandatory, prominent platform labels — rendered on the video frame itself, not buried in a three-dot menu, whenever C2PA or SynthID flags synthetic generation.
  • Algorithmic guardrails on demographic targeting — auditing recommendation loops that disproportionately funnel unlabeled synthetic personas to 55+ cohorts.
  • Adult media-literacy campaigns — aimed at retirees and older consumers rather than just school curricula.
  • Platform liability for unlabeled synthetic fraud — holding ad networks and social feeds accountable when they monetize known-synthetic impersonation scams.

Yang Mun made $300,000 in 90 days before most people knew what Veo or HeyGen could do. The next hundred Yang Muns are already rendering in the cloud.