Share:
Celebrities today. Your mom, your daughter, your colleague tomorrow. New search data on 94 famous women shows the manufacturing of fake adult imagery has gone private — and the law can’t see inside.
OnlyGuider is the world’s leading OnlyFans search engine. We index the adult creator economy all the time, which means we run regular search-data analysis of global trends across adult content — what people are searching for, where, and what that signals about where the market is going. Our Q1 2026 US analysis showed US OnlyFans search demand falling 28.5% year-on-year across all 50 states while US AI porn search demand rose 76% in twelve months.
That curve raised an obvious question: if Americans are switching from subscription content to AI-generated content at this scale, whose face are they generating? This is the answer.
TL;DR
- Taylor Swift is still America’s #1 target for synthetic adult content — 23,250 monthly searches across the deepfake and AI porn keyword family — a year on from the TAKE IT DOWN Act that her January 2024 deepfake incident helped drive into law.
- Sydney Sweeney AI porn searches: +375% YoY. Brooke Monk: +230%. Millie Bobby Brown six-month acceleration: +176%.
- The supply side is exploding. Searches for “ai onlyfans” are up +182% YoY. Searches for “ai influencer” are up +883% YoY.
- The vocabulary changed because the behaviour changed. Twelve months ago Americans typed “deepfake [name].” Today they type “[name] ai porn.” A deepfake was something you found online. AI porn is now something anyone can make at home, on a laptop, in minutes, with consumer tools promoted on social media.
- Nearly 20% of all US AI porn searches are now from people trying to make synthetic porn, not just consume it. OnlyGuider’s keyword-intent analysis classifies 19.2% of US AI porn search volume as CREATE-intent — roughly 429,000 searches per month from Americans looking for the tools to manufacture this content themselves. The TAKE IT DOWN Act criminalises publication. It says nothing about private creation.
- Celebrities are the test population. Private people are next. Same toolchain, smaller search footprint. Whoever is on the celebrity production line today is a preview of who is on the personal production line tomorrow.
What’s actually happening — in plain English
Twelve months ago, an American typing a sexual search about a celebrity into Google or Bing typed “deepfake [name].” Today they type “[name] ai porn.” That’s not just a vocabulary change. It’s a behaviour change.
A deepfake was something you found. Someone else made it, posted it to a forum, and you went looking for it. The supply chain ran through anonymous creators, image boards, and a small number of dedicated subreddits. Most users were consumers, not producers.
AI porn is something you make. The tools are now consumer-grade — face-swap apps, “undress” generators, prompt libraries, all advertised openly on TikTok, X and YouTube Shorts in the form of how-to tutorials. The whole pipeline runs on a laptop in under a minute. The user has stopped being a consumer and become a producer.
The supply-side numbers from our own keyword analysis are the receipts. US searches for “ai onlyfans” are up +182% year-on-year. US searches for “ai influencer” are up +883% year-on-year. And when we ran intent classification across the full US AI porn search corpus, 19.2% of all queries — nearly one in five, roughly 429,000 a month — came back as CREATE-intent: people looking for the tools, prompts, generators and tutorials to make this content themselves, not for pre-made content to watch. That single percentage is the demand-side driver of every name-level number in this article.
We tracked 94 named women across the full deepfake and AI porn search vocabulary, looking at Q1 2026 versus Q1 2025. Of the 64 with full year-on-year data on both sides: 57 saw their deepfake searches fall, 46 saw their AI porn searches rise, and 39 hit the textbook pattern — deepfake down, AI porn up — in the same quarter. 60 of 64 showed AI porn growth running ahead of deepfake decline.
TikTok, X and Instagram are flooded with how-to tutorials — including examples like this using Billie Eilish as the live demo target for an AI deepfake workflow that get viewed by millions of people:
The audience that was looking for fake nudes of celebrities a year ago hasn’t gone anywhere. They’ve stopped Googling and started generating. And the law passed last May to stop them only triggers the moment they upload what they’ve made — not the moment they make it.
Why this should terrify everyone, not just celebrities
This is the part of the story that doesn’t make the gossip pages.
Celebrities are the test population. They’re the names with enough public photographs in circulation for image models to produce convincing output, the names worth typing into a search bar, the names with PR cycles loud enough to drive a measurable spike in synthetic-content demand. They’re the front of the queue.
The technology doesn’t stop at celebrities. The same face-swap workflow that puts Sydney Sweeney in a generated nude works on a class photo, a LinkedIn headshot, a wedding album, an Instagram story. The 21-year-old in Austin we wrote about last week — the operator running a fictional AI OnlyFans creator pulling $43,000 a month — built her face from scratch. Same toolchain, real face, is the subject of this article. Same toolchain, your sister’s face, is the subject of nothing yet, because private people don’t generate measurable search volume. They generate evidence on someone’s hard drive.
Read this one statistic again, because it is the entire story: OnlyGuider’s analysis of US AI porn search behaviour finds that 19.2% — nearly one in every five — of every AI porn query typed into a search engine in America is from someone looking to create it, not consume it. Roughly 429,000 of these queries every single month. Roughly five million a year. Five million American attempts a year to find the tools, prompts, generators, face-swap utilities and tutorials needed to manufacture sexual imagery of a named real person at home, in private, on consumer hardware.
The celebrity numbers below are the leading indicator. The lagging indicator is going to be every classroom, every workplace, every dating app. The TAKE IT DOWN Act watches the front door of the public internet. The factory floor — what gets generated and kept on private devices — is unsupervised.
Taylor Swift, a year after her own law
A year ago, explicit AI-generated images of Taylor Swift went viral on X, reached 47 million views before the platform took action, and triggered the political momentum behind the TAKE IT DOWN Act, which President Trump signed into federal law in May 2025.
A year on, she is still America’s most-searched named woman for synthetic adult content.
The 25 most-deepfaked celebrities in America 2026
| # | Celeb | Synth searches/mo | Deepfake YoY | AI porn YoY |
|---|---|---|---|---|
| 1 | Taylor Swift | 23,250 | -46.6% | -47.9% |
| 2 | Sabrina Carpenter | 17,890 | -28.7% | +149.4% |
| 3 | Jenna Ortega | 16,580 | -33.3% | +39.3% |
| 4 | Brooke Monk | 15,960 | -41.6% | +229.6% |
| 5 | Billie Eilish | 15,300 | -21.6% | +46.4% |
| 6 | Sydney Sweeney | 12,600 | +15.9% | +375.0% |
| 7 | Millie Bobby Brown | 11,860 | +9.0% | +93.5% |
| 8 | Scarlett Johansson | 11,060 | -46.8% | -35.7% |
| 9 | Emma Watson | 7,920 | -28.7% | +8.8% |
| 10 | AOC | 7,610 | -52.5% | +0.9% |
| 11 | Addison Rae | 7,330 | -24.8% | +133.3% |
| 12 | Ariana Grande | 6,730 | -45.7% | 0.0% |
| 13 | Ice Spice | 5,730 | -41.1% | +133.3% |
| 14 | Elizabeth Olsen | 5,470 | -41.5% | +11.1% |
| 15 | Olivia Rodrigo | 4,700 | -42.9% | +64.7% |
| 16 | IU | 4,450 | -52.3% | +100.0% |
| 17 | Karina (aespa) | 4,090 | -43.1% | 0.0% |
| 18 | Natalie Portman | 3,780 | -16.4% | +118.2% |
| 19 | Gal Gadot | 3,650 | -51.7% | +33.3% |
| 20 | Zendaya | 3,350 | -23.4% | -8.6% |
| 21 | Blackpink | 3,320 | -25.9% | +100.0% |
| 22 | Margot Robbie | 3,290 | -14.3% | +85.7% |
| 23 | Charli D’Amelio | 3,010 | -47.9% | +162.5% |
| 24 | Megan Fox | 2,910 | +13.8% | +153.3% |
| 25 | Megan Thee Stallion | 2,800 | -23.6% | +42.9% |
(YoY %’s are Q1 2026 vs Q1 2025)
She is also the only real woman in our entire 75-name human dataset (94 total entities once you add fictional characters) whose deepfake AND AI porn searches both fell year-on-year. Either the law produced a direct enforcement effect in her specific case, or the platforms and image hosts have learned to nuke her content before it surfaces, knowing she is the single name in America most likely to trigger a federal takedown. Either way she stands alone as the only name where the law appears to have worked at the search level.
For the other 73 women in the dataset, demand has not fallen. It has changed shape.
Hollywood’s three compounders — Sweeney, Fox & Bieber
Three A-listers stand alone in the data: their deepfake searches AND their AI porn searches are both still rising. Every other major name in the dataset is shifting from one to the other. These three are growing on both.
Sydney Sweeney has had the loudest commercial year of any actress in America. The American Eagle “Great Jeans” campaign became a national eugenics-versus-cancel-culture row. Christy and The Housemaid shipped in November and December 2025. April 2026 brought the “Syd for Short” sequel campaign deliberately winking at the previous year’s controversy. Q1 2026: deepfake +15.9%, AI porn +375.0%. By the time she becomes a federal policy conversation, she’ll already be on 4,000+ monthly AI porn searches.
Megan Fox has not headlined a major studio film in years. She had no 2025 PR blitz on the Sweeney scale. She has no Q1 2026 release driving fresh discovery. Q1 2026: deepfake +13.8%, AI porn +153.3%. This is the cleanest demonstration in the dataset of the structural point. Image models trained on a decade of Transformers-era red-carpet footage, Jennifer’s Body publicity stills and magazine shoots do not require Megan Fox to currently be making movies. They require the reference imagery to exist. It does, in volume, and the demand sits on top of it whether or not she is culturally active.
Hailey Bieber is the most extreme case in the entire 94-entity dataset. Q1 2026: deepfake +15.4%, AI porn +600.0% — a six-fold jump in twelve months. The PR engine is well-documented: husband Justin Bieber’s public mental-health crisis through 2025 sat at the centre of entertainment coverage for nine consecutive months, and her own Rhode Beauty company sold to Elf Beauty in May 2025 for $1 billion. Every female celebrity with a tabloid-spotlight year and no major creative release this year — this is what their curve looks like.
For every actress under 35 staring down their first major franchise — and the reference-imagery archive that goes with it — the Megan Fox curve is what their AI porn footprint will look like in 2034 even if they stop working tomorrow. The asset doesn’t depreciate. It compounds.
The Top 20 Film & TV Celebs by AI porn growth YoY
| # | Actress | Deepfake YoY | AI porn YoY |
|---|---|---|---|
| 1 | Sydney Sweeney | +15.9% | +375.0% |
| 2 | Anya Taylor-Joy | -64.1% | +300.0% |
| 3 | Megan Fox | +13.8% | +153.3% |
| 4 | Salma Hayek | -51.0% | +120.0% |
| 5 | Natalie Portman | -16.4% | +118.2% |
| 6 | Millie Bobby Brown | +9.0% | +93.5% |
| 7 | Margot Robbie | -14.3% | +85.7% |
| 8 | Eva Mendes | -39.0% | +40.0% |
| 9 | Jenna Ortega | -33.3% | +39.3% |
| 10 | Gal Gadot | -51.7% | +33.3% |
| 11 | Angelina Jolie | -32.4% | +22.2% |
| 12 | Jennifer Lawrence | -14.5% | +13.6% |
| 13 | Elizabeth Olsen | -41.5% | +11.1% |
| 14 | Emma Watson | -28.7% | +8.8% |
| 15 | Hailee Steinfeld | -20.4% | -12.5% |
| 16 | Mila Kunis | -20.8% | -17.6% |
| 17 | Emma Stone | -13.7% | -21.7% |
| 18 | Zendaya | -23.4% | -8.6% |
| 19 | Scarlett Johansson | -46.8% | -35.7% |
| 20 | Penélope Cruz | -13.3% | -75.0% |
(YoY %’s are Q1 2026 vs Q1 2025)
TikTok beats Hollywood
Brooke Monk has over 40 million TikTok followers, no feature film, no album, and more synthetic-content search demand than Sydney Sweeney, Scarlett Johansson or Emma Watson. 15,960 monthly searches. AI porn up +229.6% year-on-year, +66% in the last six months.
August 2025 delivered her viral catalyst — her very public breakup with TikTok star Sam Dezz after their engagement, billed as “TikTok’s gold standard couple,” sent her engagement metrics through the roof and her name across every gossip account on the platform.
The bigger story here is structural. TikTok creators generate more reference video per unit of public profile than movie stars do, by an order of magnitude — every 30-second vertical clip is a fresh angle, a fresh expression, a fresh light setup for an image model to learn from. The result is a creator-economy class of women in their early twenties whose synthetic-content footprint already exceeds A-list actresses and is growing three times faster.
Addison Rae sits at #11 in the master ranking with 7,330 monthly searches and AI porn +133.3% YoY. Charli D’Amelio is at #23 with 3,010 monthly searches and AI porn +162.5% YoY. Three of the four largest year-on-year AI porn growth rates in the absolute-volume top 25 are TikTok-native creators. The next platform-level content-moderation fight isn’t in Hollywood. It’s on the creator lists at TikTok, Meta and YouTube.
The new music wave — pop, K-pop, and the global shift
Pop’s biggest 2025 names sit near the top of the AI porn growth ranking.
Sabrina Carpenter AI porn searches up +149.4% off the back of the Short n’ Sweet tour and album cycle. Olivia Rodrigo +64.7%. Ariana Grande, by contrast, sits at flat 0.0% AI porn YoY despite a #12 absolute-volume position — a reminder that Wicked‘s younger, female-skewing audience drives a different downstream search pattern than mainstream pop. Megan Thee Stallion +42.9%. Cardi B +125.0%. Doja Cat +75.0%. Demi Lovato +500%.
But the wave the US press hasn’t covered yet is K-pop, and the Q1 2026 numbers show it landing.
Twice as a group: AI porn +400.0% YoY. Aespa as a group: +200.0%. IU: +100%, with deepfake searches falling 52.3% in the same quarter. Blackpink as a group: +100%. (NewJeans’s contract status remains in legal flux through Q1 2026 — their YoY collapse reflects that, not demand.) Members surface individually: Lisa (Manobal) +60%, Jennie +66.7%, Karina, plus Twice members Nayeon, Jeongyeon, Sana, Mina, Momo, Tzuyu, Dahyun, Jihyo and Chaeyoung.
Jisoo of Blackpink is the first woman in our entire dataset whose AI porn search volume (1,600/mo) exceeds her deepfake search volume (980/mo). She is what the end-state of this looks like — a generation of named targets who skip the “deepfake era” entirely because the technology rebranded before they got famous in the West.
By the end of 2026 we expect at least three more K-pop names to cross the threshold into AI-porn-dominant search, and Twice to break into the absolute-volume top 25 alongside her.
The Top 28 Music Celebs by AI porn growth YoY
| # | Artist | Deepfake YoY | AI porn YoY |
|---|---|---|---|
| 1 | Demi Lovato | -12.2% | +500.0% |
| 2 | Twice (K-pop) | -43.7% | +400.0% |
| 3 | Aespa (K-pop) | -46.2% | +200.0% |
| 4 | Sabrina Carpenter | -28.7% | +149.4% |
| 5 | Ice Spice | -41.1% | +133.3% |
| 6 | Cardi B | -1.4% | +125.0% |
| 7 | IU | -52.3% | +100.0% |
| 8 | Blackpink | -25.9% | +100.0% |
| 9 | Shakira | +6.1% | +100.0% |
| 10 | Doja Cat | -32.8% | +75.0% |
| 11 | Jennie (Blackpink) | -31.9% | +66.7% |
| 12 | Olivia Rodrigo | -42.9% | +64.7% |
| 13 | Lisa (Blackpink, Manobal) | -4.8% | +60.0% |
| 14 | Camila Cabello | -42.1% | +50.0% |
| 15 | Dua Lipa | -34.7% | +50.0% |
| 16 | Billie Eilish | -21.6% | +46.4% |
| 17 | Megan Thee Stallion | -23.6% | +42.9% |
| 18 | Jisoo (Blackpink) | -22.3% | +33.3% |
| 19 | Miley Cyrus | -52.2% | +31.8% |
| 20 | Katy Perry | -55.3% | +6.2% |
| 21 | Beyoncé | -63.7% | +5.6% |
| 22 | Taylor Swift | -46.6% | -47.9% |
| 23 | Selena Gomez | -53.7% | -25.0% |
| 24 | Lana Del Rey | -37.5% | -40.0% |
| 25 | NewJeans (K-pop) | -67.4% | -57.1% |
| 26 | Lady Gaga | -13.6% | 0.0% |
| 27 | Rihanna | -47.1% | 0.0% |
| 28 | Ariana Grande | -45.7% | 0.0% |
(YoY %’s are Q1 2026 vs Q1 2025)
Models, reality TV & business — every named figure with year-on-year data, ranked by AI porn growth
Figures whose primary professional identity is modelling, reality TV, or beauty/lifestyle business rather than acting or music. Listed separately so readers can read each table cleanly.
| # | Name | Deepfake YoY | AI porn YoY |
|---|---|---|---|
| 1 | Hailey Bieber (Rhode Beauty / model) | +15.4% | +600.0% |
| 2 | Bella Hadid (model) | -31.0% | +100.0% |
| 3 | Kim Kardashian (SKIMS / reality TV) | -18.6% | +61.9% |
| 4 | Kendall Jenner (model / reality TV) | -16.7% | +57.1% |
| 5 | Kylie Jenner (Kylie Cosmetics / reality TV) | -9.6% | +23.1% |
(YoY %’s are Q1 2026 vs Q1 2025)
The fictional women — and the actresses they drag with them
Nineteen fictional female characters surfaced in our dataset with measurable monthly synthetic-content search demand. The Final Fantasy VII character Tifa Lockhart pulls 760 monthly searches — more than Kim Kardashian’s pure synthetic-content footprint. Catwoman’s AI porn searches are up 400% year-on-year. Wonder Woman is up 70%. Lara Croft is up 140%.
The pattern that matters: fictional characters are a leading indicator for the live actresses who play them. The image models love fictional characters because the reference imagery is perfect — official 3D models, decades of canonical art, infinite fan-art training data, no rights-holder takedown leverage on the underlying character design. And the audience that searches for them feeds straight back into the actresses attached.
| # | Character | Maps to | Deepfake YoY | AI porn YoY |
|---|---|---|---|---|
| 1 | Catwoman | Zoë Kravitz / Anne Hathaway / Michelle Pfeiffer | 0.0% | +400.0% |
| 2 | Samus | Metroid | n/a | +350.0% |
| 3 | Aerith | Final Fantasy VII | -50.0% | +200.0% |
| 4 | Bayonetta | Bayonetta franchise | n/a | +200.0% |
| 5 | Lara Croft | Angelina Jolie / Alicia Vikander | +16.7% | +140.0% |
| 6 | Wonder Woman | Gal Gadot | -36.4% | +70.0% |
| 7 | Princess Peach | Nintendo (Anya Taylor-Joy, voice) | 0.0% | +70.0% |
| 8 | Mystique | X-Men | n/a | +50.0% |
| 9 | Black Widow | Scarlett Johansson | -51.0% | +47.1% |
| 10 | Jinx | Arcane / League of Legends | -100.0% | +42.9% |
| 11 | Princess Leia | Carrie Fisher | -25.6% | +11.1% |
| 12 | Tifa Lockhart | Final Fantasy VII | -44.4% | +7.1% |
| 13 | Harley Quinn | Margot Robbie / Lady Gaga | -16.7% | +7.1% |
| 14 | Zelda | Nintendo | NEW | +6.2% |
| 15 | Chun-Li | Street Fighter | -33.3% | 0.0% |
| 16 | Rias | High School DxD anime | n/a | 0.0% |
| 17 | Rogue | X-Men | n/a | -66.7% |
(YoY %’s are Q1 2026 vs Q1 2025)
Catwoman +400% in the year before The Batman Part II ships in October 2027 is not a coincidence. Wonder Woman +70% with no greenlit film in production and no actress currently attached is fan demand bleeding back to Gal Gadot because the demand has nowhere else to go. The gaming and Marvel character cluster is rotating into AI porn faster than most live actresses, and every time a fictional character moves, a real woman’s curve moves with her.
Women in Politics — and the international outlier
US politicians are smaller absolute targets than entertainers, but the same shift is visible.
AOC’s Q1 2026 deepfake searches dropped 52.5% YoY — the steepest decline of any name in our top 25 by volume — while her AI porn searches stayed essentially flat at +0.9%. Kamala Harris: AI porn -56.7% on a 1,810/mo base, the largest collapse in the politics cluster. Hillary Clinton: 80/mo, AI porn +60%, deepfake +15.2%. Melania Trump: 270/mo, AI porn +42.9%. Sarah Palin: 50/mo, AI porn +100%.
The outlier sits outside the United States.
Italian Prime Minister Giorgia Meloni has 24,150 sexualised monthly searches, 569 distinct adult-keyword variants, and an absolute synthetic-content demand footprint that beats Taylor Swift’s. On a per-capita basis she beats Swift by roughly six times. Italian deepfake searches for Meloni are up 17% YoY; Italian searches for “giorgia meloni nuda” are up 35%. A sitting head of government is a larger named synthetic-content target than the most-searched pop star in America. Full Italian breakdown is in our Italy OnlyFans vs AI Porn Q1 2026 analysis.
The hidden cases — Kim Kardashian and Jennifer Lawrence
Kim Kardashian’s name will be conspicuous by its near-absence from the volume table above. She is in the data — she just sits at #38 by synthetic-content footprint, not #1.
Why? Because her synthetic-content footprint (1,600/mo) is dwarfed by her leaked-photo footprint — searches for the 2007 sex tape and the historical material — at 39,500/mo. That’s a pre-AI demand pool we deliberately exclude from the synthetic-content column. Folding it in would corrupt the read on what’s specifically AI-driven.
When you isolate her synthetic-content numbers alone, she is rotating exactly like everyone else: deepfake -18.6% YoY, AI porn +61.9% YoY, net positive movement of +81 percentage points.
Jennifer Lawrence is the same case at greater scale. Her leaked-photo footprint is 122,230/mo — the lingering demand from the 2014 iCloud breach. Her synthetic-content footprint is 1,430/mo, with deepfake -14.5% and AI porn +13.6%. The shift is real but slow. Both Kardashian and Lawrence will eventually look like everyone else in this dataset. The shift just has more historical weight to drag.
The 25 fastest-growing Celebrity AI porn targets in America
For journalists tracking momentum rather than absolute size — the names whose AI porn search demand grew fastest in Q1 2026 vs Q1 2025:
| Rank | Celeb | AI porn YoY |
|---|---|---|
| 1 | Hailey Bieber | +600.0% |
| 2 | Demi Lovato | +500.0% |
| 3 | Twice | +400.0% |
| 4 | Sydney Sweeney | +375.0% |
| 5 | Anya Taylor-Joy | +300.0% |
| 6 | Brooke Monk | +229.6% |
| 7 | Aespa | +200.0% |
| 8 | Charli D’Amelio | +162.5% |
| 9 | Megan Fox | +153.3% |
| 10 | Sabrina Carpenter | +149.4% |
| 11 | Addison Rae | +133.3% |
| 12 | Ice Spice | +133.3% |
| 13 | Cardi B | +125.0% |
| 14 | Salma Hayek | +120.0% |
| 15 | Natalie Portman | +118.2% |
| 16 | IU | +100.0% |
| 17 | Blackpink | +100.0% |
| 18 | Shakira | +100.0% |
| 19 | Millie Bobby Brown | +93.5% |
| 20 | Margot Robbie | +85.7% |
| 21 | Doja Cat | +75.0% |
| 22 | Wonder Woman (fictional) | +70.0% |
| 23 | Princess Peach (fictional) | +70.0% |
| 24 | Jennie | +66.7% |
| 25 | Kim Kardashian | +61.9% |
(YoY %’s are Q1 2026 vs Q1 2025)
A handful of these are small absolute bases throwing big percentages — Bieber’s six-fold jump comes off 220 monthly searches, not 22,000 — but every one of them is a leading-indicator name. The absolute volumes catch up. The direction is what matters.
The law is slow.
Here is the policy implication of every number above.
The TAKE IT DOWN Act criminalises the publication and distribution of non-consensual AI-generated intimate imagery. It triggers when a victim files a takedown. The 48-hour clock starts when something goes public.
It has nothing to say about the act of creation itself — the AI porn sitting on a private device, generated for personal use, never uploaded.
429,000 monthly US searches are now from people looking for the tools to make this content. Five million attempts a year, by US users alone, to manufacture sexual imagery of named real people in private. The supply-side numbers we cited at the top of this piece — “ai onlyfans” +182% YoY, “ai influencer” +883% YoY, the open how-to tutorial culture on TikTok, X and YouTube Shorts — describe a toolchain that is now consumer-grade and a demand curve for it that is exploding.
The law watches the front door — what gets posted in public. Millions of Americans are now operating a back room — what gets generated in private. The front door is monitored. The back room is unsupervised. The data above shows whose faces are coming off the production line.
The celebrity production line is the visible one because the search demand is large enough to measure. The private production line is the invisible one — and there is no reason to believe the same toolchain that generated five million celebrity attempts last year is not also being used on people whose names don’t show up in keyword research because they aren’t famous enough to search for.
Celebrities today. Family members tomorrow. The shift from “deepfake [name]” to “[name] ai porn” — public consumption to private creation — is the most consequential change in the celebrity-harm economy since the term “deepfake” was coined in 2017. It is happening in plain sight, in the search data, right now. The law passed to stop it triggers on the door, not the factory floor.
Methodology
This analysis builds on OnlyGuider’s Q1 2026 US OnlyFans search analysis, which found US OnlyFans search demand falling 28.5% year-on-year while AI porn search demand rose 76% in twelve months. This piece answers the follow-on question: whose face is the AI porn demand pointed at?
We pulled US search data on 94 named entities — 75 real women, 19 fictional female characters — across four keyword families: “deepfake [name],” “[name] ai porn,” “[name] nudes,” “[name] leaked.” Absolute monthly volumes come from Semrush US keyword cluster data (March 2026). Deepfake year-on-year comes from Google Ads API , comparing Q1 2025 to Q1 2026 monthly averages. AI porn year-on-year comes from Bing Ads API, same Q1 windows — because Google does not return search volume for phrases containing “ai porn” under its adult-content advertising policy.
The “synth searches per month” column in our tables is deepfake plus AI porn volume only. It deliberately excludes leaked-photo and OnlyFans-related search volume, which are pre-AI demand pools that would distort the read on the synthetic-content economy. .
Copyright & Attribution
The America’s Most Deepfaked Celebrities Report 2026 © 2026 by OnlyGuider is licensed under CC BY 4.0. Link to data source required
Author
Want more paying OnlyFans subscribers?
Launch a campaign in 24 hours and fill your creators’ pages with real fans, not freebie hunters.
Recent partner average: $4.24 back per $1 spent • 324% ROMI
Real people only. No bots, no spam traffic sources.
ONLY FOR ONLYFANS AGENCIES, MODELS, AND MARKETERS
Want to promote your OnlyFans profile?
Start getting paying subscribers from search. Launch in 48 hours.
Recent partner average: $4.24 back per $1 • 324% ROMI