Deepfake porn is sexual imagery created or altered with AI to show a real person who never consented. 98% of the 95,820 deepfake videos found online in 2023 were pornographic, and 99% of the people targeted were women, according to Security Hero, a company that reviews identity-protection products.
It doesn’t only happen to celebrities. 2.2% of 16,693 adults surveyed in 10 countries said they had been victims of deepfake porn, according to Umbach et al. (2024).
In this article, we’ll break down what deepfake porn is, how common it is, who is targeted, how big the “nudify” economy has become, how it affects victims, where it’s illegal and how to get it removed.
Top Deepfake Porn Statistics
- 98% of deepfake videos found online in 2023 were pornographic, up from 96% in 2019.
- 99% of people targeted in deepfake porn are women, and 94% work in the entertainment industry.
- 2.2% of adults across 10 countries say they’ve been victims of deepfake porn, and 1.8% admit perpetrating it.
- 181 nudify websites averaged more than 40.1 million unique visitors a month between December 2025 and March 2026.
- 46 women in the European Parliament, about 1 in 6 female MEPs, were found on deepfake porn sites in 2026.
- 26.1% of UK adults fear being targeted by deepfake porn, rising to 33.2% of women.
- 46 US states criminalize sexual deepfakes of adults, and federal law requires platforms to remove reported images within 48 hours.
What Is Deepfake Porn?
Deepfake porn is any sexual image or video made or edited with AI or other software so that it appears to show a real, identifiable person without their consent. 96% of deepfake videos online were pornographic when the detection company Deeptrace counted them in 2019, according to CNN, and 98% were in 2023, per Security Hero.
It usually takes one of two forms: a face swapped onto another person’s body in a video, or an ordinary photo digitally “undressed” by a nudify tool. Both are image-based sexual abuse, because the person depicted never agreed to it.
US federal law calls these images “digital forgeries,” meaning intimate images created or altered with software, machine learning or AI that look authentic to a reasonable person, according to Cornell Law School.
Here’s a breakdown of the two main counts of deepfake porn online:
| Report | Deepfake Videos Counted | Share Pornographic | Porn Targets Who Are Women |
|---|---|---|---|
| Deeptrace (2019) | 14,678 | 96% | 100% |
| Security Hero (2023) | 95,820 | 98% | 99% |
Both counts come from cybersecurity companies, as FactCheckNI notes, though it rated the 99% figure accurate. They remain the latest available counts (2023).
How Common Is Deepfake Porn?
Deepfake videos online grew about 6.5 times in four years, from 14,678 in 2019 to 95,820 in 2023, based on counts by Deeptrace and Security Hero (our calculation). Deepfake porn videos alone rose 464%, from 3,725 in 2022 to 21,019 in 2023, according to Security Hero.
Most of it sits on dedicated sites. Dedicated deepfake porn sites host 90% of deepfake porn videos, and the top 10 had 303,640,207 total video views and 34,836,914 monthly visits in 2023, per Security Hero.
One of the biggest deepfake porn marketplaces hosted 42,986 sexual deepfake videos of 3,803 people, with over 1.5 billion views, according to a study by Han et al. (2024). The site shut down in May 2025 after a critical service provider ended its service, according to UC Berkeley School of Information.
Here’s a breakdown of how much deepfake porn is online:
| Measure | Figure | Source |
|---|---|---|
| Deepfake videos online, December 2018 | 7,964 | Deeptrace |
| Deepfake videos online, 2019 | 14,678 | Deeptrace |
| Deepfake videos online, 2023 | 95,820 | Security Hero |
| Deepfake porn videos, 2022 | 3,725 | Security Hero |
| Deepfake porn videos, 2023 | 21,019 | Security Hero |
| Share of deepfake porn on dedicated sites (2023) | 90% | Security Hero |
| Video views on the top 10 dedicated sites (2023) | 303,640,207 | Security Hero |
| Videos on one of the biggest marketplaces | 42,986 | Han et al. (2024) |
| Views on that marketplace | Over 1.5 billion | Han et al. (2024) |
Many people have seen it. 43% of UK people aged 16+ saw at least one deepfake in the six months to June 2024, and 14% of adults who had seen deepfakes said they had seen a sexual one, according to Ofcom. Separately, 18.8% of 1,403 UK adults said they had been exposed to pornographic deepfakes, per the Alan Turing Institute (2024).
In a company poll of 1,522 US men who had watched porn in the past six months, 48% said they had seen deepfake porn at least once (Security Hero, 2023).
How Many People Are Victims of Deepfake Porn?
2.2% of adults in 10 countries reported that someone had created, shared or threatened to share sexual deepfakes of them, and 1.8% admitted doing it themselves, in a survey of 16,693 adults in May and June 2023 by Umbach et al., presented at CHI 2024.
Victimization was highest in Australia (3.7%), South Korea (3.1%) and Mexico (2.9%), according to the full paper. Perpetration peaked in the US at 2.6% and was lowest in Belgium at 0.3%, so self-reported perpetration in the US was more than 8 times Belgium’s.
Newer surveys find higher rates. 7.0% of 7,230 adults in Australia, the UK and the US reported AI-generated image-based sexual abuse in a 2024 survey, and 15.1% reported abuse involving digitally altered images of any kind, according to Henry et al. (2026).
The rates differ partly because of what they count. Henry’s 15.1% includes non-AI editing, and a 2019 survey of people aged 16 to 84 in the UK, Australia and New Zealand found 14.1% on a similarly broad measure, according to Monash University.
Here’s a breakdown of deepfake porn victimization and perpetration in surveys:
| Population | Measure | Rate | Survey |
|---|---|---|---|
| Adults in AU, UK and US | Any digitally altered sexual image | 15.1% | Henry et al. (2024 fieldwork) |
| Ages 16–84 in UK, AU and NZ | Any digitally altered sexual image | 14.1% | Flynn et al. (2019 fieldwork) |
| Adults in AU, UK and US | AI-generated sexual image | 7.0% | Henry et al. (2024 fieldwork) |
| Adults in Australia | Deepfake porn victim | 3.7% | Umbach et al. (2023 fieldwork) |
| Adults in South Korea | Deepfake porn victim | 3.1% | Umbach et al. |
| Adults in Mexico | Deepfake porn victim | 2.9% | Umbach et al. |
| Adults in 10 countries | Deepfake porn victim | 2.2% | Umbach et al. |
| Adults in 10 countries | Perpetrated deepfake porn | 1.8% | Umbach et al. |
| Adults in the US | Perpetrated deepfake porn | 2.6% | Umbach et al. |
Who Is Targeted by Deepfake Porn?
Women, overwhelmingly. 99% of the people targeted in deepfake porn online in 2023 were women, and 94% of those depicted worked in the entertainment industry, according to Security Hero. In 2019, all of the pornographic deepfakes Deeptrace found featured women, per CNN.
South Korean singers and actresses make up 53% of the people featured, followed by Americans at 20%, per Security Hero. One deepfake image of Taylor Swift was viewed more than 47 million times before X removed it in January 2024, according to Monash University.
Private individuals are targeted too. On one of the biggest marketplaces, actors made up 58% of targets, but 14% of the 1,942 targets researchers checked fell below the site’s own celebrity thresholds, according to Han et al. Among UK adults who had seen a sexual deepfake, 64% said it showed a celebrity or public figure, 15% someone they know and 6% themselves, per Ofcom (2024).
Politicians are frequent targets. 26 members of Congress, 25 of them women, were depicted in sexual deepfakes, nearly 1 in 6 congresswomen, according to the American Sunlight Project, as reported by The Markup (2024). In October 2026, an audit of 200 websites found 46 women and 4 men from the European Parliament’s 720 members, about 1 in 6 female MEPs, according to Agora Digitale Transformation, as reported by Anadolu Agency.
Here’s a breakdown of how many deepfake porn targets are women:
| Group | Women’s Share | Source |
|---|---|---|
| People targeted in deepfake porn online (2023) | 99% | Security Hero |
| Deepfake victims supported in South Korea (1,616 people, 2025) | 97.8% | Korean government data |
| Members of Congress depicted (2024) | 96.2% (25 of 26) | American Sunlight Project |
| Most common targets on one large marketplace | About 95% | Han et al. |
| MEPs found on deepfake porn sites (2026) | 92% (46 of 50) | Agora Digitale Transformation |
Population surveys tell a more mixed story. In the 10-country survey, 1.6% of men and 0.7% of women said someone had created sexual deepfakes of them, a relative risk of 2.23 for men, per Umbach et al. Henry et al. also found higher victimization among men, but women reported a broader range of harms.
The data measure different things, and neither study explains the gap: content counts show who appears on public sites, mostly famous women, while surveys capture self-reported private experiences, including threats.
How Big Is the Nudify App Problem?
181 nudify websites averaged more than 40.1 million unique visitors a month between December 2025 and March 2026, according to the Institute for Strategic Dialogue (ISD). In September 2023, Graphika counted over 24 million unique visitors to 34 such providers.
That’s about 1.7 times as many visitors, though the two studies tracked different sets of sites (our calculation). The 16 most-visited sites alone drew over 200 million visits in the first six months of 2024, according to the San Francisco City Attorney.
Social platforms sent about 5.79 million visits to the 10 most-visited nudify sites from December 2025 to February 2026, led by YouTube (1.82 million) and X (1.3 million), per ISD, and links advertising these services on X and Reddit rose more than 2,000% in 2023, per Graphika.
X’s Grok chatbot generated an estimated 3 million sexualised images in 11 days, according to a January 2026 estimate by the Center for Countering Digital Hate, as reported by CP24. The estimate did not say how many images were made without the consent of the people shown.
Here’s a breakdown of the nudify economy:
| Measure | Figure | Source |
|---|---|---|
| Unique visitors to 34 nudify providers, September 2023 | 24 million+ | Graphika |
| Rise in links advertising nudify services on X and Reddit, 2023 | 2,000%+ | Graphika |
| Visits to the 16 most-visited sites, January–June 2024 | 200 million+ | San Francisco City Attorney |
| Average monthly unique visitors, 181 sites, December 2025–March 2026 | 40.1 million+ | ISD |
| Visits referred by social media to the top 10 sites, December 2025–February 2026 | About 5.79 million | ISD |
| Estimated yearly revenue of AI nudifier sites | Up to $36 million | Indicator (2025) |
| Sexualised images generated by Grok in 11 days (estimate) | About 3 million | CCDH (2026) |
How Does Deepfake Porn Affect Victims?
There’s little survey data on harms specific to deepfakes, but fear of being targeted is widespread. 26.1% of UK adults fear being targeted by deepfake porn, including 33.2% of women and 18.2% of men, so women are about 1.8 times as likely to be afraid, according to the Alan Turing Institute (2024).
The harm falls unevenly. Women reported a broader range of harms and greater negative emotional and social impacts from AI-generated sexual images, according to Henry et al. Across all forms of image-based sexual abuse, real or fake, 33.4% of victim-survivors in 10 countries reported mental health harms and 61.4% at least one harm, according to Umbach, Henry and Beard.
Here’s a breakdown of attitudes toward deepfake porn:
| Measure | Figure | Population |
|---|---|---|
| Concerned about deepfake porn as violence against women and girls | 89.6% | UK adults (2024) |
| Parents who want nudifying tools banned for everyone | 80% | 2,000 UK parents (2024) |
| Back strict laws against harmful deepfakes | 72.1% | UK adults (2024) |
| Fear being targeted by deepfake porn | 26.1% | UK adults (2024) |
| Men who had seen deepfake porn and don’t feel guilty (company poll) | 74% | US men who watch porn (2023) |
| Would want to report it if someone close became a victim (company poll) | 73% | US men who watch porn (2023) |
Helpline figures are still small. The UK Revenge Porn Helpline logged 63 cases involving synthetic sexual images in 2025, up from 58 in 2024, which the helpline says likely understates the real scale, according to SWGfL.
Is Deepfake Porn Illegal?
In the US, yes. Since May 19, 2025, the federal TAKE IT DOWN Act has made knowingly publishing a sexual deepfake of an adult without consent a crime punishable by up to 2 years in prison, and threatening to publish one up to 18 months, according to Cornell Law School.
46 states also criminalize sexual deepfakes of adults, and 35 of them added that coverage in 2024 or 2025, in our review of state codes. Ballotpedia counts 48 states with laws addressing sexually explicit deepfakes as of August 2026; in our review, that broader count includes laws limited to minors or civil lawsuits.
Penalties range from misdemeanors to felonies, and in most states deepfakes fall under each state’s revenge porn law.
Prosecutions have started. The first person believed to be convicted under the federal law pleaded guilty on April 7, 2026, according to the DOJ. On May 20, 2026, two men in Brooklyn were charged with allegedly posting deepfake porn of about 90 and about 50 women, per the DOJ.
England and Wales made sharing intimate images, including deepfakes, with intent to cause distress or for sexual gratification, or threatening to share them, punishable by up to 2 years in prison from January 31, 2024, per legislation.gov.uk, and creating or requesting a sexual deepfake of an adult became a crime on February 6, 2026, under the Data (Use and Access) Act 2025, per legislation.gov.uk.
South Korea has seen a sharp rise in cases. Reported deepfake sex crimes rose from 31 in 2020 to 550 in 2024, more than 17 times as many, according to The Korea Times, and police arrested 963 people in a seven-month crackdown, per Korea JoongAng Daily.
Here’s a breakdown of deepfake porn laws and enforcement:
| Jurisdiction | Law or Action | Key Figure |
|---|---|---|
| US (federal) | TAKE IT DOWN Act (2025) | Up to 2 years in prison; 48-hour platform removal |
| US (states) | State deepfake laws | 46 states criminalize adult deepfakes (48 per Ballotpedia, August 2026) |
| US (federal) | DOJ domain seizure (June 2026) | 2 deepfake sites seized |
| England and Wales | Sharing or threatening to share (since January 31, 2024) | Up to 2 years in prison (sharing with intent, or threatening) |
| England and Wales | Creating or requesting a deepfake (since February 6, 2026) | New criminal offence |
| South Korea | Possessing or watching sexual deepfakes (2024) | Up to 3 years in prison |
| South Korea | Deepfake-related crimes, November 2024–October 2025 | 1,553 cases |
| Australia | Online Safety Act civil penalty (2025) | A$343,500 for deepfakes of 6 women |
| European Union | Directive 2024/1385 | Member states must criminalise by June 14, 2027 |
How Are Platforms and Regulators Responding?
10 of the most-visited nudify websites are now offline or inaccessible in California, after San Francisco sued the operators of 16 sites in August 2024, according to the San Francisco City Attorney (June 2025). One operator agreed to a permanent injunction and $100,000 in civil penalties.
In June 2025, Meta sued a Hong Kong-based developer that kept advertising nudify apps after its ads were removed, and it had shared more than 3,800 URLs of such services with other companies, according to Meta. Google says ranking changes cut exposure to explicit fake images in name searches by over 70%, per Google (2024).
Ofcom opened an investigation into X over Grok imagery in January 2026, with possible fines of up to 10% of worldwide revenue, per Ofcom, and the FTC sent warning letters to 12 nudify companies on May 20, 2026, per the FTC.
Here’s a breakdown of recent actions against deepfake porn:
| Who | Action | Key Figure |
|---|---|---|
| San Francisco City Attorney | Lawsuit against nudify sites (2024–2025) | 10 of 16 sites offline or inaccessible in California |
| Meta | Lawsuit and link-sharing (2025) | 3,800+ URLs shared; 4 ad networks disrupted |
| Search ranking changes (2024) | Over 70% less exposure to explicit fakes | |
| FTC | Warning letters (May 2026) | 12 nudify companies |
| State attorneys general | Letter to xAI over Grok (January 2026) | 35 attorneys general |
| Ofcom (UK) | Investigation into X over Grok (January 2026) | Fines up to 10% of worldwide revenue |
| European Commission | Digital Services Act investigation into X (January 26, 2026) | Formal investigation opened |
| eSafety Commissioner (Australia) | Enforcement against nudify services (2025–2026) | 7 most widely used services withdrew or complied |
In Australia, three nudify services that drew about 100,000 visits a month from Australian users withdrew from the country after eSafety enforcement action, which could have led to penalties of up to A$49.5 million, according to eSafety. The services later returned under new ownership with age checks, according to a later eSafety release.
How to Get Deepfake Porn Removed
Removal usually works when victims report. Over 94% of images reported to the UK Revenge Porn Helpline for 2025 and earlier cases had been removed, according to SWGfL, and Australia’s eSafety Commissioner had material removed in 89.9% of its removal requests from 2018 to 2023, per eSafety.
These are the main routes, in order:
- Save evidence. Screenshot the images and record links, usernames and dates before asking for removal.
- Block re-uploads with StopNCII.org. StopNCII accepts AI-generated nude or sexual images, creates a digital fingerprint on your device so the image never leaves it, and shares only that fingerprint with partner platforms. It is for people who were 18 or older in the image, and it was protecting 2 million images by November 2025, up 97% from 2024, with 17 partner platforms, per SWGfL.
- Ask Google to remove search results. Google removes fake sexual images of you from Search, per Google, and aims to review requests within 48 hours, per its help page.
- Report it to the platform. US platforms covered by the TAKE IT DOWN Act must remove reported images within 48 hours of a valid request, per Cornell Law School. If they don’t, you can complain to the FTC at TakeItDown.ftc.gov, according to the FTC.
- If the image shows someone under 18, use NCMEC’s Take It Down service instead.
Deepfakes are handled through the same channels as real intimate images, so the full steps for reporting revenge porn, including helplines in the US, UK, Canada and Ireland, apply here too.
Methodology
We reviewed 158 data points from 48 sources. Content counts come from two companies, Deeptrace (2019, reported by CNN) and Security Hero (2023), which publishes identity-protection product reviews and earns affiliate commissions; both analysed videos on dedicated deepfake sites and video platforms. The marketplace figures come from a measurement study of one site by Han et al. (2024).
Victimization rates come from online surveys of adults: Umbach et al. (16,693 adults in 10 countries, May–June 2023), Henry et al. (7,230 adults in Australia, the UK and the US, July–August 2024) and Flynn et al. (UK, Australia and New Zealand, 2019). Exposure and attitude data come from Ofcom (2,000 UK people aged 16+, June 2024), the Alan Turing Institute (1,403 UK adults, May–June 2024) and Internet Matters (2,000 UK parents, June 2024). The Security Hero survey of 1,522 US men is a company poll.
Surveys use different definitions: some count only AI-generated images, others any digitally altered image, and “victimization” includes creation, sharing and threats. Nudify traffic figures come from Graphika, ISD and the San Francisco City Attorney and cover different sets of sites. The count of 46 states comes from our review of state criminal codes. Multipliers, ratios and the women’s shares for Congress and MEPs are our own calculations.
Key Takeaways
Deepfake porn is AI-made or AI-altered sexual imagery of real people without consent, and it makes up 98% of deepfake videos online (2023). 99% of the people targeted are women, mostly celebrities, though 2.2% of adults in 10 countries say it has happened to them.
The tools are spreading fast, with 181 nudify sites drawing more than 40.1 million unique visitors a month by early 2026. Laws are catching up: 46 US states and federal law now criminalize it, platforms must remove reported images within 48 hours, and over 94% of images reported to the UK helpline were removed.
FAQ
What is deepfake porn?
Deepfake porn is sexual imagery created or altered with AI to depict a real person who didn’t consent, usually a face swap or a digitally “undressed” photo. 98% of deepfake videos found online in 2023 were pornographic, according to Security Hero.
How common is deepfake porn?
Very common online. Security Hero counted 95,820 deepfake videos in 2023, about 6.5 times the 14,678 found in 2019. In surveys, 2.2% of adults in 10 countries say they have been victims, and 7.0% of adults in Australia, the UK and the US reported AI-generated sexual image abuse.
Who is targeted by deepfake porn?
Mostly women. 99% of people targeted in deepfake porn online are women, and 94% work in entertainment. Private individuals are targeted too: 15% of UK adults who saw a sexual deepfake said it showed someone they know.
Is deepfake porn illegal?
In many places, yes. In the US, publishing a sexual deepfake of an adult carries up to 2 years in prison under federal law, and 46 states criminalize it. England and Wales made creating one a crime on February 6, 2026.
How do I get deepfake porn removed?
Report it to the platform, which in the US must remove it within 48 hours of a valid request, and ask Google to remove search results. StopNCII.org accepts AI-generated images from adults and was protecting 2 million images by November 2025.
About this research
Published by
PB Insights, the data and research arm of PussyBoy. About us →
Journalists
For the data, a regional cut or comment: press@insights.pussyboy.net
Reuse
Licensed CC BY 4.0. Cite "PB Insights" and link to this report.
"According to adult-industry data analysts at PB Insights…"
