A $975 million asset transfer. Ads that were never filmed airing on American television. A Chinese grandmother talking every week to her deceased son. Behind these seemingly unrelated stories, a single movement is taking shape: the cloning of creators by artificial intelligence is becoming industrialized, and audience trust is beginning to crack.
From experiment to industry
Until recently, AI clones of influencers looked like fringe experiments. That has changed.
In the United States, fake pro-Trump influencer profiles, entirely generated by artificial intelligence, were created in India by scammers targeting conservative communities. During the last Coachella festival, accounts simulated their attendance with fake selfies, fake outfits, fake backdrops, and a VIP lifestyle fabricated from scratch.
The nature of the use has shifted. AI is no longer used to retouch content, it manufactures identities, lifestyle scenes, and false proof of influence. Creators themselves are beginning to test digital doubles capable of handling part of their activity: responding, engaging, producing, selling, staying visible around the clock. The appeal is obvious: less time on camera, more monetization. The movement remains fragmented, sometimes makeshift. But it points to one reality: identity manufacturing has become a commodity like any other.
Khaby Lame, or the mirage of the digital twin
In January 2026, Khaby Lame suddenly thrust the issue into the spotlight. The most-followed creator on TikTok, with 161 million subscribers, announced the sale of his image management company, Step Distinctive Limited, to Hong Kong-listed group Rich Sparkle Holdings. Stated value: $975 million.
The industrial ambition was unprecedented: leveraging the creator’s image, voice, and biometric data to build a “digital twin” dedicated to live shopping, with a stated target of several billion dollars in annual sales. The deal ultimately unraveled, with serious doubts raised about the financial structure, but the underlying logic holds. A creator’s value is no longer measured by audience size alone. It now includes what their image and voice can produce in an automated, parallel, infinite manner. YouTube has already deployed a “likeness detection” tool allowing creators and public figures to identify AI-generated videos using their likeness without authorization.
Tibo InShape and iShowSpeed: when identity escapes its owner
Not all creators are willing participants in the commercialization of their image. Many are victims of it. In October 2025, French influencer Tibo InShape allowed users to generate cameos of him on Sora, OpenAI’s consumer application. The experiment quickly spiraled out of control: offensive and overtly racist videos circulated widely, to the creator’s dismay. We cover this episode in detail in a video available on our website.
Streamer iShowSpeed (54M+ followers on YouTube) experienced an even more brutal version of the same problem. He discovered an ad he had never filmed, for a “blockchain casino” brand he had never heard of aired in the middle of an official NBA game. AI fraud is no longer confined to fake accounts and shady news feeds. It now infiltrates established media environments, exploits their credibility, and exposes broadcasters, ad networks, platforms, advertisers, and creators alike. The audience of an official sports broadcast can, without knowing it, be exposed to fraudulent ads linked to a brand that will never find out.
Why this concerns everyone
The most unsettling case had nothing to do with marketing. In Shandong Province, China, the only son of a woman in her eighties dies in a road accident. Fearing that the news could be fatal for his elderly grandmother with a heart condition, her grandson contacts a provider specializing in AI clones. He supplies photos, videos, and audio recordings. The company generates an avatar capable of speaking and reproducing the deceased’s personality traits. The old woman continues to call her son by video, unaware that he is gone.
This proliferation is feeding a deeper mistrust. According to an IFOP study, only 6% of French people say they are certain they can detect AI-generated content.
For brands, the risk does not only come from what they publish. It comes from what others can publish in their name, using their logo, their identity, their executives, their spokespeople. This is no longer an influencer problem. It is a trust problem affecting everything that circulates online.
What regulation opens, certification locks in
Europe is putting a first guardrail in place. From August 2, 2026, Article 50 of the AI Act will make transparency obligations enforceable: providers and deployers will be required to ensure that synthetic content is detectable in a machine-readable format, and to clearly disclose any use of deepfakes. This is an essential step. It is not complete protection.
Where the AI Act requires disclosure of what is synthetic, certification makes it possible to prove what is authentic. The two approaches are complementary. The question is no longer only about detecting what is false, it is about demonstrating what is true. Content certification anchors proof of prior creation and authenticity at the very moment of publication, enforceable in the event of identity theft. This is the reflex that most online communications still lack, and the only one that makes a digital identity truly defensible.
Discover how Certiphy.io protects your digital identity →
In the age of clones, your digital identity is only yours if you can prove it.
