An AI influencer is a social media persona whose face, body, voice and personality are generated by software rather than belonging to a real person. It holds accounts on Instagram, TikTok or YouTube, publishes posts on a schedule, replies to comments and signs brand deals in the same way a human creator does. A team behind it - an agency, a studio, a brand or a single operator - writes the character, generates the content and collects the fees. The format exists because a synthetic creator never ages, never cancels a shoot, can appear anywhere for the cost of compute and cannot embarrass a sponsor with an off-script remark. It also exists because audiences follow it, which is the part regulators and platforms spent 2025 and 2026 trying to make transparent.

How an AI influencer is produced and sold

An AI influencer starts as a character brief: a name, an age, a city, a niche such as fitness or fashion, and a backstory written to make the account read like a life. Early personas were built in 3D modelling software and rendered frame by frame. Most current ones use generative image and video models, typically adapted on a set of reference images so that the same face appears consistently across hundreds of posts. Voice is synthesised separately, and a language model may draft captions and direct-message replies.

The commercial mechanics then mirror a human brand deal. A brand or its agency briefs the operator, agrees deliverables and a fee, and the operator produces the posts. Revenue also comes from subscription platforms and affiliate links. Business Insider reported in 2023 that Aitana López, a persona run by the Barcelona agency The Clueless, earned up to $11,000 in a month, while other outlets have cited figures from EUR 4,000 to about $17,000. Those numbers trace back to the agency and are not audited.

The difference from a human deal sits in the labelling layer. As of October 2026, Instagram asks the operator to switch on an "AI-generated profile" attribute in profile settings; TikTok asks for its AI-generated content (AIGC) label on realistic posts; YouTube asks for the "altered content" setting in YouTube Studio.

Origins: from CGI characters to generative personas

The modern format dates to April 23, 2016, when an Instagram account called Miquela published its first post. Created by Trevor McFedries and Sara DeCou, the character was rendered in computer-generated imagery (CGI) and presented without explanation. In April 2018 a staged "hack" storyline revealed that Miquela and a rival character were both products of Brud, a Los Angeles start-up. Time listed Miquela among its 25 most influential people on the internet in June 2018. NFT company Dapper Labs bought Brud on October 4, 2021, according to TechCrunch, which said the start-up had reached a $125 million valuation.

Photographer Cameron-James Wilson created Shudu, a "digital supermodel", in April 2017, and drew criticism because a white man had built a Black female persona. A May 2019 Calvin Klein advert pairing Miquela with model Bella Hadid led to an apology.

Generative AI changed the economics from 2023. A persona no longer needed a 3D artist for every image. Meta launched 28 AI characters on September 27, 2023, and opened AI Studio to US users on July 29, 2024. H&M said on March 28, 2025 that 30 models had agreed to the creation of AI "digital twins", and a Guess advert in the August 2025 print edition of Vogue used an AI-generated model disclosed only in small print.

Disclosure and labelling rules

United States. The Federal Trade Commission (FTC) approved revised Endorsement Guides on June 29, 2023, effective July 26, 2023. The new definition says an endorser "could be or appear to be an individual, group, or institution." Responding to commenters, the Commission said the language was not limited to virtual influencers but also covered fake reviewers and "non-existent entities that purport to give endorsements." A synthetic creator promoting a third-party product therefore owes the same disclosure as a human one. New York went further: bill S8420A, signed on December 11, 2025, requires advertisers with actual knowledge to conspicuously disclose a synthetic performer from June 9, 2026, with fines of $1,000 for a first violation and $5,000 thereafter.

India and China. The Advertising Standards Council of India issued influencer guidelines on June 1, 2021, effective June 14, requiring a virtual influencer to "disclose to consumers that they are not interacting with a real human being." China's labelling measures, issued on March 14, 2025 by the Cyberspace Administration of China and three other agencies, took effect on September 1, 2025, and require visible and metadata labels on AI-generated images, audio and video.

European Union. Article 50 of the AI Act became applicable on August 2, 2026. Article 50(4) obliges deployers to disclose deep fakes, and Article 50(1) requires that people be told when they are interacting with an AI system, which reaches personas that chat with followers. The lighter regime for "evidently artistic, creative, satirical, fictional" work is narrow: the Commission guidelines of July 20, 2026 treat synthetic humans persuading viewers to buy a product as fully in scope. Fines reach EUR 15 million or 3% of worldwide turnover.

Industry codes. IAB's AI Transparency and Disclosure Framework, first released in January 2026, lists synthetic avatars that simulate human interaction among uses needing a label. Version 2, issued on August 18, 2026, put it plainly: "A photorealistic AI influencer run as paid social and display creative requires disclosure. An animated gecko mascot does not," according to PPC Land's report.

Platform policies as of October 2026

Instagram began testing an opt-in "AI creator" label in May 2026. On August 31, 2026 it renamed it "AI-generated profile" and, according to its Creators blog, warned that operators who do not label "may see limits to their profile's reach". Detected unlabelled accounts become ineligible for recommendations until they add the label or win an appeal. Instagram conceded that "our systems might not always be perfect."

TikTok introduced its creator AIGC label on September 19, 2023, and said on May 9, 2024 that more than 37 million creators had used it; it also reads Coalition for Content Provenance and Authenticity (C2PA) metadata to auto-label imported content. YouTube made disclosure of realistic synthetic content mandatory on May 21, 2025, and in May 2026 moved those labels into more visible positions. Snap stopped recommending wholly AI-generated videos on Spotlight in July 2026, and Spotify introduced an "AI Persona" badge on August 11, 2026 that removes badged artists from default recommendations.

Why marketers pay attention

The appeal is control. The Clueless, founded by designer Rubén Cruz, built Aitana because it was tired of human influencers "who have egos", according to Business Insider. A persona can be localised, scaled and kept on brand indefinitely.

The cost is trust and suitability. In Integral Ad Science's survey of 290 US media professionals, 82% said creator brand safety would matter and 83% said rising AI-generated content on social media needed monitoring. IAB found 39% of Gen Z consumers felt negative about AI ads, against 20% of Millennials, and an NYU Stern finding cited in the IAB framework showed that disclosing generative AI cut click-through by 31.5%.

Limitations and disputes

Research questions effectiveness. A University of Florida study by Fanjue Liu and Yu-Hao Lee, published in the Journal of Product & Brand Management in February 2024, found virtual influencers were "less effective endorsers than human influencers due to a lower perception of authenticity," though comparable for functional products.

Deception is the sharper problem. The New York Times reported in March 2026 on AI personas, including an "Amish" woman and a "monk", selling supplements across Instagram, TikTok and Facebook without clear disclosure. Representation is another: Shudu in 2017 and Meta's own AI profiles, deleted in January 2025 after a backlash over a persona called Liv, drew criticism for synthetic identities built by people outside the groups depicted. Labour groups object too. SAG-AFTRA said on September 30, 2025 that the AI "actress" Tilly Norwood "is not an actor, it's a character generated by a computer program."

Legal gaps remain. Lawyers at InfoLawGroup argued in August 2023 that FTC guidance leaves room for brand-owned personas to be treated as spokespersons rather than endorsers. Market-size estimates for the sector vary widely and come mostly from vendors.

What an AI influencer is not

  • AI avatar of a real creator. A human's licensed digital likeness, such as YouTube's personal avatars for Shorts or H&M's digital twins, represents an existing person. Right of publicity law and consent govern it.
  • Creator who uses AI tools. A human influencer editing with AI is not an AI influencer; Instagram exempts such accounts from its profile label.
  • Virtual influencer. Largely a synonym, but the older term covers CGI characters such as early Miquela that predate generative AI.
  • Deepfake. Synthetic media impersonating a real person, usually without consent. An AI influencer is an invented person.

Recent developments

Meta stopped new character creation in AI Studio on August 10, 2026. YouTube extended likeness detection to every creator aged 18 and over on May 16, 2026, and TikTok Shop's May 23, 2026 rules ban AI voices standing in for live presenters. Denmark tabled bill L 49 on October 8, 2026, requiring consent for realistic digital imitations from January 1, 2027. The next EU milestone is December 2, 2026, when generative AI providers already on the market must bring machine-readable marking into conformity.

Timeline

  • April 23, 2016 - Miquela's first Instagram post.
  • April 2017 - Cameron-James Wilson creates Shudu.
  • April 2018 - Brud revealed as Miquela's creator.
  • June 2018 - Time names Miquela among the 25 most influential people on the internet.
  • May 2019 - Calvin Klein apologises over its Miquela and Bella Hadid advert.
  • June 1, 2021 - ASCI guidelines require virtual influencers to disclose they are not human (effective June 14).
  • October 4, 2021 - Dapper Labs acquires Brud.
  • June 29, 2023 - FTC approves revised Endorsement Guides (effective July 26).
  • September 19, 2023 - TikTok launches its creator AIGC label.
  • September 27, 2023 - Meta introduces 28 AI characters.
  • November 2023 - Business Insider reports Aitana López's earnings.
  • July 29, 2024 - Meta opens AI Studio to US users.
  • January 2025 - Meta deletes its AI character profiles after backlash.
  • March 28, 2025 - H&M announces 30 digital twins of models.
  • May 21, 2025 - YouTube's synthetic content disclosure becomes mandatory.
  • September 1, 2025 - China's AI labelling measures take effect.
  • September 30, 2025 - SAG-AFTRA statement on Tilly Norwood.
  • December 11, 2025 - New York signs synthetic performer law S8420A.
  • January 2026 - IAB releases AI Transparency and Disclosure Framework.
  • May 2026 - Instagram tests "AI creator" label.
  • June 9, 2026 - New York synthetic performer disclosure takes effect.
  • August 2, 2026 - EU AI Act Article 50 becomes applicable.
  • August 11, 2026 - Spotify announces AI Persona badges.
  • August 18, 2026 - IAB framework version 2 names photorealistic AI influencers.
  • August 31, 2026 - Instagram launches "AI-generated profile" label with reach limits.
  • October 8, 2026 - Denmark tables deepfake bill L 49.
  • December 2, 2026 - EU marking conformity deadline for existing generative AI systems.

Summary

Who. Agencies, studios, brands and individual operators that build and run synthetic personas; the platforms that host them, led by Instagram, TikTok and YouTube; and regulators including the FTC, New York State, ASCI, China's CAC and the European Commission.

What. A social media creator that does not exist: a generated face, voice and personality that posts, gathers followers and endorses products, distinct from a real creator's AI avatar or a human who uses AI tools.

When. The CGI era began with Miquela in April 2016; generative AI made personas cheap from 2023; disclosure duties tightened between June 2023 and August 2026, with Instagram's reach penalty arriving on August 31, 2026.

Where. Mainly on Instagram, TikTok and YouTube, with rules set in the US, New York, India, China and the EU, and Denmark's bill pending.

Why. Brands gain control, scale and consistency. Audiences and regulators want to know whether the person recommending a product is real, and the evidence so far suggests disclosure lowers engagement, which is why labelling is contested.