Do you have to show your face to sell on TikTok and Instagram?

The short answer

Below: what the rules name, what the research measured (and didn't), why hands are the best-evidenced faceless format, a table of six ways to make a video without your face, and the one line an AI or hired presenter can't cross.

What the platforms' rules actually name

We read the current rule pages rather than summaries of them. Faces appear in none of them as a requirement or a signal. Originality appears in almost all of them.

One detail in Instagram's April 2026 post is worth dwelling on: it says 75% of recommendations in the US now come from original posts. That's a first-party number, and it's the clearest statement either platform has made about where recommendation reach goes. It says nothing about faces. It says a great deal about reposting.

The face research, read properly

The "faces get more engagement" claim almost always traces back to one paper. It's a real paper and a good one — it just answers a narrower question than it gets quoted for.

Why the two face studies disagree

The 2014 paper, from Georgia Tech and Yahoo researchers, analysed 1.1 million Instagram photos and found those with faces were 38% more likely to receive likes and 32% more likely to receive comments. It's correlational, and its authors said plainly that they stopped short of explaining why. The 2020 Journal of Marketing Research paper by Li and Xie did something the first didn't: it corrected for selection bias — the fact that the kinds of posts that include faces may differ in other ways from those that don't. After that correction, faces raised engagement on Twitter but not on Instagram. The explanation offered is that faces, selfies especially, are so common on Instagram that one stops being distinctive.

The one finding that held across both platforms and both product categories in the 2020 study wasn't about people at all: high-quality, professionally shot images consistently earned more engagement. For a store deciding whether to get on camera, that's the more transferable result. Craft travelled; faces didn't.

Hands: the finding that actually matters to a store

In 2022 the Journal of Marketing Research published eight studies by Luangrath, Peck, Hedgcock and Xu on what they call vicarious touch: watching a hand in contact with a product on a screen. Using images, GIFs and virtual reality, they found it increased psychological ownership — the sense of "that's mine" — and how much people valued the product. They named it the vicarious haptic effect.

Two conditions make it genuinely useful rather than a platitude:

1. The hand has to be touching the product. The authors state it isn't enough to show a hand in the ad. A hand gesturing near a product, or pointing at a caption, doesn't do it.
2. The movement has to be diagnostic. The effect weakened when the interaction shown wasn't a diagnostic hand movement. Our reading of that in practical terms: show the hand doing what a shopper would do in a shop to judge the thing — squeezing the cushion, pulling the zip, pouring from the jug, stretching the fabric, closing the clasp.

Caveat, stated honestly: the studies used images, GIFs and VR in controlled settings, not TikTok feeds, and they measured valuation, not checkout. It's still the nearest peer-reviewed evidence to "does this help sell the product", and it points at a format that costs a phone and a table. It also suits what Instagram says it ranks on — a satisfying, finishable demonstration is the kind of thing people watch to the end. More on building one in how to make a product video that actually gets watched.

Six ways to make a video without your face

Five of these are legitimate. One looks like the cheapest and is the one the platforms have written rules against.

Why supplier videos deserve their own warning

This is the failure faceless store accounts are actually exposed to, and it's easy to walk into with permission in hand. Instagram's April 2026 post states its originality guidelines are separate from its intellectual property policies. So a supplier saying "use our videos" settles the copyright question and leaves the originality one untouched. TikTok puts it in one line: reused material "without anything new" is ineligible for the For You feed.

Instagram is also specific about what doesn't count as adding something: "just adding a border, watermark, subtitles, or a credit in the captions." What does count, in its words, is additions that "meaningfully enhance" the content — such as "unique text on the content itself that provides more context than just describing what's happening". In practice: if you can't re-shoot, re-edit around a question the supplier's clip doesn't answer. If the account has already been flagged, Instagram's Account Status screen shows it, and eligibility returns once most of what you've posted over a rolling 30 days is original.

Choosing a format by product, not by preference

The useful question isn't "am I comfortable on camera". It's "what does a buyer need to see to judge this". The flow below is our judgement, built from the rules and research above.

The one line a hired or AI presenter can't cross

Faceless doesn't have to mean personless, and once there's a person on screen who isn't you, a different set of rules applies. They're simple, and they're the same whether the presenter is paid or generated.

The FTC's rule on reviews and testimonials, finalised in August 2024, bans testimonials that misrepresent that they are "by someone who does not exist, such as AI-generated fake reviews, or who did not have actual experience with the business or its products". Read against video, that draws a clean line: a presenter can describe a product — what it's made of, what it does, what it costs. A synthetic presenter can't say "I've used this every morning for a month", because that person doesn't exist. A paid creator can say it only if it's true, and the FTC's Endorsement Guides require the paid relationship to be disclosed clearly.

AI presenters also need the platform label: TikTok requires "clear labeling when AI or editing is used to realistically depict people", and Meta requires disclosure on photorealistic AI video. Neither platform treats the format as illegitimate — TikTok itself launched Symphony avatars for advertisers in June 2024, stock avatars "created using paid actors that are licensed for commercial use". The label is the price of entry, not a penalty; the details are in does AI content get penalised on TikTok and Instagram.

And if the footage in question belongs to your customers — a photo someone tagged you in — that's a rights question rather than an originality one, covered in can you repost a customer's photo.

Where Magik fits, and where it doesn't

Magik makes one of the formats in the table: AI creator videos, a presenter-style clip generated from the product photos already in your Shopify store, posted to TikTok, Instagram and Facebook on a schedule. Because each one is built from your own catalogue rather than reposted from somewhere else, it sits on the right side of both originality rules above. It needs the AI label, like any realistic synthetic presenter — and the rule to hold any AI presenter to, ours included, is the one above: it describes the product; it doesn't claim to have used it.

What it can't do is the thing the best evidence on this page points at. A generated video can't show your hand pulling your zip, and it can't show your dress on a real body. For products that live or die on those, film the hands-only demo or pay a creator — and use generated video for the volume around them, since the number of attempts is what a small account is short of. If you're weighing tools that make presenter videos, the comparison page covers Xyla AI, Minta, Outfy, IDEQO and Buffer alongside us.

Frequently asked

Sources

Platform rules were read on the platforms' own pages in September 2026. Research is cited to the peer-reviewed papers, with their limits stated in the text.

1. Instagram for Creators — Rewarding original creators on Instagram, 30 April 2026. Source of the extension of originality rules to photos and carousels; that accounts that "primarily post unoriginal content" will "no longer be shown in places where we recommend content"; the rolling 30-day eligibility and Account Status; the definition of original content as "work that you wholly created or reflects your unique perspective"; what does and doesn't qualify; the statement that the rules are separate from IP policies; and "75% of recommendations in the US now coming from original posts".
2. TikTok — Community Guidelines, Integrity and Authenticity. Source of "Content is also ineligible for the FYF if it includes unoriginal or reused material without anything new", the FYF-ineligible example of "clips that show someone else's watermark or logo", and "We require clear labeling when AI or editing is used to realistically depict people or scenes".
3. Instagram — Instagram Ranking Explained, 31 May 2023. Source of the four Reels ranking predictions.
4. Bakhshi, Shamma & Gilbert — "Faces engage us: photos with faces attract more likes and comments on Instagram", CHI 2014; figures as reported by the Georgia Tech News Center. Source of the 1.1 million photos, the 38% likes and 32% comments figures, and the authors' statement that they stopped short of determining why.
5. Li & Xie — "Is a Picture Worth a Thousand Words? An Empirical Study of Image Content and Social Media Engagement", Journal of Marketing Research, 2020. Source of the selection-bias correction, the face effect on Twitter but not Instagram, the airline and SUV categories, and the consistent effect of high-quality, professionally shot images. The explanation that selfies make faces too common on Instagram to stand out is as summarised by Baylor University's Keller Center.
6. Luangrath, Peck, Hedgcock & Xu — "Observing Product Touch: The Vicarious Haptic Effect in Digital Marketing and Virtual Reality", Journal of Marketing Research, 2022. Source of the eight studies, the effect on psychological ownership and valuation, "it is not enough to show a hand in an advertisement; the hand must be touching a product", and attenuation when the movement is not diagnostic. The practical examples of diagnostic movements are our interpretation.
7. FTC — Final rule banning fake reviews and testimonials, 14 August 2024. Source of the ban on testimonials "by someone who does not exist, such as AI-generated fake reviews, or who did not have actual experience with the business or its products".
8. FTC — Updated Endorsement Guides, June 2023. Source of the requirement to disclose material connections such as payment.
9. TikTok Newsroom — Announcing Symphony Avatars, 17 June 2024. Source of stock avatars "created using paid actors that are licensed for commercial use".
10. Meta — Our approach to labeling AI-generated content and manipulated media, April 2024. Source of the disclosure requirement for photorealistic AI video and realistic audio.
11. Deliberately not cited: vendor claims that faceless accounts convert better or worse than face-on-camera ones; reach-change percentages attributed to Instagram's 2026 originality update by third-party blogs (Instagram's own post gives none); and "faceless niche" earnings figures. None disclosed a sample or linked to a first-party source. The formats table and the product flow are our judgement built from the sources above.

Keep reading