Brands are quietly deploying AI-generated influencers on social media to promote their products, with no legal requirement to disclose that the people in the ads aren’t real. A Guardian investigation found that 40-60% of content from major brands may be AI-made, with some creators signing non-disclosure agreements to prevent the practice becoming public.
🔍 THE BOTTOM LINE
The influencer economy — already built on artifice — is being replaced by something cheaper and more controllable: synthetic personas that never sleep, never age, and never disagree with the brand. There are no disclosure rules in New Zealand. The EU’s AI Act will require labelling from August 2026, but it won’t apply here. The Advertising Standards Authority in the UK has confirmed it has “nothing in our rules” that forces brands to reveal when an influencer is fabricated. The same gap exists in NZ.
What the Investigation Found
The Guardian’s Sarah Marsh identified multiple brands using AI-generated influencers on Instagram. Once, a photo app, posted videos of a bride crying about using its disposable camera app at her wedding. The woman was almost certainly not real. Maket, an AI housing design platform, posted a video of a woman who appears to be AI-generated saying she “could kiss the interior designer” who showed her the app. Ashle, a Dubai-based fashion brand, posted photos of a woman wearing its clothes at a restaurant — with a visible extra finger. After the Guardian asked about it, the photos were deleted.
Clarissa Mansbridge, a former celebrity manager who creates AI influencers for brands through her Mia Metaverse portfolio, told the Guardian that 40-60% of content from major brands may be AI-made. “If you sign with a brand, they’ll make you sign an NDA saying you can’t talk about the fact they’re using AI, because consumer trust is still being built,” she said. “I call it plausible deniability.”
The Regulatory Gap
The UK’s Advertising Standards Authority confirmed there are no specific rules requiring AI disclosure. “There’s nothing in our rules that prohibits this and there are no disclosure rules for AI content labelling,” an ASA spokesperson said. The regulator would only act if the ad itself is misleading — not if the person in it doesn’t exist.
The EU’s AI Act will begin requiring labels on AI-generated or manipulated content — including deepfake images, audio, and video — from August 2026. But the UK left the EU, and New Zealand has no equivalent legislation on the table.
Which? conducted a separate investigation finding that 70% of people cannot correctly identify all deepfake videos shown to them. Lisa Barber, Which? Tech editor, said: “It is concerning that consumers are not able to trust the content they are seeing. Companies must be transparent when content has been created using AI.”
NZ Angle
New Zealand has no specific AI disclosure requirement for advertising. The Advertising Standards Authority NZ’s code requires ads be “legal, decent, honest and truthful” — but there’s no explicit provision for AI-generated personas. The Privacy Commissioner has flagged AI as a priority area but has not issued guidance on synthetic influencers specifically.
The gap matters more for a small market like NZ. Global brands running AI influencer campaigns on Instagram and TikTok don’t geo-fence their content. A Kiwi scrolling Instagram sees the same fabricated bride, the same synthetic fashion model, as someone in London. With no local rule and no EU-style labelling on the horizon, the disclosure burden falls entirely on the brand — and they have no incentive to be honest.
This connects to the broader pattern we’ve been tracking: Signal’s Meredith Whittaker warning that agentic AI is a surveillance backdoor, and Granta’s decision to stop publishing prize winners because AI broke literary trust. The thread is the same — the gap between what AI can do and what the public knows it’s doing is widening, and the institutions meant to police it are behind.
The Other Side
Brands using AI influencers argue the practice is no different from using professional actors or models. Maket told the Guardian that AI influencers were “one of several ways for us to test creative concepts and marketing hooks at a small scale before investing in broader campaigns.” Ashle said its early marketing imagery “utilised AI during our initial launch phase to showcase designs” and that the garments themselves are real.
The argument has a surface logic: advertising has always been fabricated. Professional models aren’t ordinary customers. Retouching is standard. But there’s a categorical difference between enhancing a real person and inventing one. A model who endorses a product can be held accountable — they can be asked whether they actually use it, whether they experienced the results they claim. An AI influencer cannot. There’s no person behind the claim, just a brand talking to itself through a synthetic face.
The Bigger Picture
The influencer marketing industry was worth $24 billion in 2024 and is projected to reach $32 billion by 2026. If 40-60% of brand content is already AI-generated, the economics are straightforward: a synthetic influencer costs a fraction of a human one, never requires a fee renegotiation, and can’t go off-brand in a livestream. The market will optimise toward fabrication unless regulators intervene.
The deeper issue is what happens to trust. Social media platforms already struggle with misinformation. When the line between a real person’s recommendation and a brand’s self-promotion disappears entirely, the concept of “social proof” — the psychological mechanism that makes influencer marketing work — breaks down. You can’t have social proof from a thing that has no society.
❓ FAQ
Is AI influencer advertising legal in NZ? Yes. There’s no specific prohibition. The ASA NZ code requires ads be “honest and truthful” but doesn’t address whether the person in the ad needs to exist. A complaint could be made under the misleading conduct provisions, but the ASA has not tested this.
How can I tell if an influencer is AI-generated? The Guardian’s investigation identified tells: extra fingers, inconsistent lighting, unnatural skin texture, and generic captions. But the technology is improving fast. Which? found 70% of people can’t identify all deepfakes — meaning you probably can’t either.
Will NZ get EU-style AI labelling rules? Not currently. The government’s AI strategy, released in 2025, focuses on enabling adoption rather than restricting it. There’s no equivalent to the EU AI Act’s deepfake labelling requirement in the legislative pipeline.
What does this mean for human influencers? The mid-tier influencer market is most at risk. Top influencers with genuine audience relationships will survive. But the thousands of creators making modest incomes from brand deals are competing with free, perfect, compliant AI avatars. The economics don’t favour humans.
