Deepfake pornography has become a growing concern as artificial intelligence tools make it easier to create realistic sexual images and videos without a person’s consent. The scale of this content has also increased sharply, with 95,820 deepfake videos identified in a 2023 study, compared with 14,678 in 2019. This rapid growth highlights how quickly synthetic sexual content has expanded alongside improvements in AI technology.
Earlier research showed a strong connection between synthetic media and pornography, while women make up the overwhelming majority of identified targets. Celebrities and entertainment figures have historically received much of the attention, but research also shows that private individuals can become victims. Deepfake pornography is distributed through dedicated websites, social media platforms, forums, messaging services, and other online channels, making it difficult to measure and remove completely.
In this article, we are going to take a look at Deepfake Porn Statistics, including its growth, who is most targeted, where it is distributed, how activity can move between platforms, the latest legal developments, and the challenges involved in detecting and removing it.
Key Deepfake Porn Statistics
- 95,820 deepfake videos were identified in a 2023 Security Hero study, compared with 14,678 in 2019.
- 98% of the 95,820 identified deepfake videos were pornographic in the 2023 study.
- 96% of 14,678 identified deepfake videos were pornographic in Deeptrace’s 2019 study.
- 99% of people targeted in identified deepfake pornography were women in the 2023 Security Hero study.
- 99% of people featured on the deepfake pornography websites studied by Deeptrace worked in entertainment, mainly as actresses or musicians.
- 2.2% of more than 16,000 people across 10 countries reported experiencing non-consensual synthetic intimate imagery.
- 94% of identified pornographic deepfake videos in Deeptrace’s 2019 study were hosted on dedicated deepfake websites, while 6% appeared on mainstream pornography websites.
- 12 nudification companies received warning letters from the U.S. Federal Trade Commission in May 2026.
Deepfake Porn Growth Over Time
Deepfake sexual content was already growing rapidly before generative AI tools became widely available. Research from Deeptrace identified 7,964 deepfake videos in December 2018. By 2019, the number had increased to 14,678, almost doubling in less than a year. Deeptrace reported that 96% of the identified videos were pornographic.
A later investigation by Security Hero identified 95,820 deepfake videos in 2023, which was about 550% higher than the 2019 figure. The study found that 98% of the identified videos were pornographic.
98% of Deepfake Videos Were Pornographic in a Major Study
Sexual content has made up a large share of the deepfake videos identified in major studies. A 2023 Security Hero study found that 98% of 95,820 identified deepfake videos were pornographic. This means that approximately 9 out of every 10 identified videos in the study were sexual in nature.
The pattern was similar in an earlier Deeptrace study. In 2019, Deeptrace found that 96% of 14,678 identified deepfake videos were pornographic.
This doesn’t mean that 98% of all AI-generated images and videos today are pornographic. Modern generative AI is used for many purposes, including entertainment, advertising, education, political content, and fraud.
Who Is Most Targeted by Deepfake Porn?
Women are the most common targets of deepfake pornography, while celebrities and other public figures have historically made up a large share of identified cases. However, research and recent law-enforcement cases show that private individuals can also be targeted.
Celebrities Are Highly Visible Targets of Deepfake Porn
Early deepfake pornography studies found that public figures, especially people working in entertainment, made up most identified targets. Deeptrace reported in 2019 that 99% of people featured on the deepfake pornography websites it analysed worked in the entertainment industry, mainly as actresses or musicians.
Security Hero reported a similar pattern in its 2023 study, with entertainment figures making up a large share of the identified content. The study also found particularly high levels of targeting involving South Korean singers and actresses.
| Metric | % Who Are Targets of Deepfake Porn |
| Entertainment workers among Deeptrace targets | 99% |
| People reporting non-consensual synthetic intimate imagery | 2.2% |
| People reporting perpetration-related behaviour | 1.8% |
Women Account for Almost All Deepfake Porn Targets
Women make up the overwhelming majority of people targeted in deepfake pornography. A Security Hero study found that 99% of people appearing in identified deepfake pornography were women. The finding has also been cited by Australia’s eSafety Commissioner as evidence of the highly gendered nature of deepfake image-based abuse.
An earlier Deeptrace study reported an even stronger pattern. 100% of the people featured on the deepfake pornography websites analysed in the study were female.
Ordinary People Are Also Victims of Deepfake Porn
Deepfake pornography is not only a celebrity issue. A peer-reviewed study involving more than 16,000 people across 10 countries found that 2.2% had personally experienced non-consensual synthetic intimate imagery. Another 1.8% reported behaviour related to creating or sharing such content.
Recent law-enforcement cases also show that private individuals can be targeted. In May 2026, the U.S. Department of Justice charged two people over the alleged distribution of AI-generated deepfake pornography involving both celebrities and non-public individuals.
The shows that celebrity deepfakes are easier for researchers to identify because public figures are recognisable and their content often attracts more attention. Deepfakes involving classmates, former partners, co-workers, or other private individuals may circulate within smaller communities and never appear in major research datasets.
Deepfake Porn Distribution by Platform Statistics
94% of Identified Deepfake Porn Was Hosted on Dedicated Deepfake Websites
Deeptrace’s 2019 study found that 94% of the pornographic deepfake videos it identified were hosted on dedicated deepfake websites. Only about 6% appeared on mainstream pornography websites.
This finding shows that, during the early development of deepfake pornography, specialised websites were the main distribution channel. However, this statistic reflects Deeptrace’s 2019 dataset and should not be treated as a current estimate of the entire internet.
6% of Identified Deepfake Porn Appeared on Mainstream Pornography Websites
Around 6% of the pornographic deepfake videos identified by Deeptrace in 2019 were found on mainstream pornography websites. The relatively small share suggests that most identified deepfake material at the time was concentrated on websites specifically created or used for deepfake content.
The figure also provides an important historical comparison because deepfake material is now distributed through a much wider range of online services.
12 Nudification Companies Received FTC Warning Letters in 2026
In May 2026, the U.S. Federal Trade Commission sent warning letters to 12 companies offering nudification tools. These services use AI to create synthetic nude or sexually explicit images from photographs.
The action shows that regulators are increasingly focusing on the creation and distribution of deepfake sexual content, rather than only the websites that host the finished material.
Deepfake Porn Is Now Distributed Across Multiple Online Channels
Although there is no reliable percentage breakdown for 2026, deepfake sexual content can now circulate through dedicated adult websites, social media platforms, image and video-sharing services, online forums, private communities, messaging apps, file-sharing services, and nudification platforms.
This wider distribution makes the ecosystem harder to measure than it was when dedicated deepfake websites accounted for most identified content.
Deepfake Porn Migration After Platform Shutdowns Statistics
Research suggests that shutting down a major deepfake pornography platform does not necessarily eliminate related activity. A 2026 study examined the shutdown of MrDeepFakes alongside the introduction of the U.S. TAKE IT DOWN Act and tracked activity across three other online forums. The researchers found increases in deepfake-related sharing and requests across the platforms studied.
3 Online Forums Were Studied After the Platform Shutdown
The study examined three online forums that contained sections dedicated to different types of porn content. Researchers compared activity in sections focused on synthetic non-consensual explicit imagery with activity in other pornography categories. This approach allowed them to measure whether activity changed after the intervention rather than simply counting posts before and after the shutdown.
Deepfake Sharing and Requests Increased Across Platforms
The researchers observed an increase in both the sharing and requesting of synthetic non-consensual explicit imagery (SNCEI) across the websites studied. In some cases, they also found an increase in the number of new contributors. The size and timing of these changes differed between platforms, showing that users did not respond to the shutdown in exactly the same way everywhere.
Platform Shutdowns Did Not Suppress Activity Overall
The study found no overall suppression of SNCEI activity following the combined regulatory and platform intervention. Instead, the researchers concluded that the activity appeared to be redistributed across platforms. This suggests that removing one major website may change where deepfake pornography is shared without necessarily reducing the underlying activity.
Deepfake Porn Laws and Regulations by Country
Governments are introducing stronger laws to address non-consensual deepfake pornography, AI-generated intimate images, and the platforms and tools used to create or distribute them. The legal approach differs by country. Some jurisdictions have created specific offences for deepfake sexual content, while others rely on broader image-based abuse, online-safety, privacy, or AI laws.
1. United States
The United States introduced one of the most significant federal measures through the TAKE IT DOWN Act, which was signed into law on May 19, 2025. The law covers non-consensual intimate imagery, including certain AI-generated or digitally altered intimate content.
From May 19, 2026, covered platforms must provide a system that allows victims to report qualifying content and request its removal. Platforms generally have 48 hours to remove validly reported intimate imagery and known identical copies. The Federal Trade Commission is responsible for enforcing these platform requirements.
Federal prosecutors have also begun using the law in deepfake-related cases. In May 2026, the U.S. Department of Justice announced charges against two individuals accused of creating and distributing AI-generated deepfake pornography involving celebrities and private individuals.
Individual U.S. states can also have separate laws covering non-consensual intimate images, sexual deepfakes, impersonation, and related forms of abuse.
2. United Kingdom
The UK has also strengthened its legal response to sexually explicit deepfakes. In 2026, the government announced legislation making it illegal to create or request intimate deepfake images of adults without their consent. The government has also stated that sharing or threatening to share intimate deepfake material without consent is already covered by criminal law and the UK’s online-safety framework.
The UK has also announced additional measures aimed at making online platforms respond more quickly to reports of non-consensual intimate imagery. These measures focus on reducing the time that harmful material remains available after it has been reported.
3. European Union
The European Union uses a combination of AI regulation and broader digital-safety rules rather than relying on a single deepfake pornography law.
The EU AI Act introduced transparency requirements for deepfakes. From August 2, 2026, applicable rules require certain AI-generated or manipulated content to be identifiable, while people exposed to deepfakes in covered situations must be informed that the material has been artificially generated or manipulated.
The EU has also moved toward directly regulating AI systems used to create non-consensual sexual content. Under the AI Omnibus, AI systems that generate non-consensual sexually explicit or intimate content, including certain “nudification” applications, are prohibited. The prohibition is scheduled to take effect in December 2026.
This means the EU’s approach covers both the content itself and the AI systems capable of generating abusive sexual material.
4. Australia
Australia addresses deepfake pornography mainly through its broader image-based abuse and online-safety framework. The eSafety Commissioner can investigate complaints involving image-based abuse and take action to have harmful material removed.
Australia has also taken enforcement action involving AI-generated sexual imagery. In 2025, an Australian Federal Court ordered a $343,500 penalty in a case involving deepfake images of Australian women.
The eSafety Commissioner has also acted against services that allow users to create synthetic nude images, including services that can be used to create sexualised images of children.
Deepfake Porn Regulations Are Expanding in 2026
The legal response to deepfake pornography is moving beyond traditional privacy and harassment laws. In 2026, the main areas of regulation include criminalising non-consensual creation and distribution, requiring platforms to remove reported material, regulating AI systems that can generate abusive sexual content, and increasing transparency around AI-generated media.
However, the laws are not uniform. A person creating or sharing a sexual deepfake may face very different legal consequences depending on the country involved. The rapid development of AI tools also means that governments continue to update their laws as new forms of synthetic sexual content emerge.
Deepfake Porn Detection and Takedown Challenges
Detecting and removing deepfake pornography is difficult because deleting one copy does not necessarily remove all versions of the content from the internet. Once an image or video is downloaded, it can be copied, edited, renamed, and uploaded to multiple websites and platforms.
1. Edited Copies Can Avoid Simple Detection Systems
Deepfake content can be modified in several ways, including cropping, filtering, compression, adding watermarks, changing the frame rate, or creating a new synthetic version from the original image. These changes can make it harder for automated systems to recognise that two files contain the same underlying content.
Hashing technology can help address this problem. StopNCII creates a digital fingerprint, or hash, of an intimate image without requiring the original image to be uploaded to the service. Participating platforms can compare new uploads against these fingerprints and take action when they identify a match.
However, StopNCII states that cropping, filtering, or editing an image can prevent the modified version from matching its original hash. In such cases, the altered image may need to be hashed separately.
2. Private and Encrypted Platforms Limit Detection
Private groups and encrypted messaging services create another major challenge. StopNCII explains that its technology works with participating public-facing platforms and does not cover most encrypted messaging services. This means that content shared privately may not be detected through the same systems used on participating public platforms.
3. AI Detection Systems Also Face Accuracy Challenges
Detecting whether an image or video has been manipulated is becoming more difficult as synthetic media becomes more realistic. Detection tools must identify altered content while avoiding false positives, where genuine images or videos are incorrectly classified as deepfakes.
This creates a continuing challenge for platforms and researchers. Detection systems need to keep improving as the techniques used to create and modify synthetic media also evolve.
4. StopNCII Supports Deepfake and Synthetic Intimate Images
StopNCII.org allows adults to create a digital fingerprint of qualifying non-consensual intimate images while keeping the original file on their device. The service explicitly includes deepfake and synthetic intimate images when the image depicts the person using the service, they have access to it, and it contains nude or semi-nude content.
The service can be used internationally, but protection depends on whether the platform where the content appears participates in the StopNCII system.
5. Take It Down Is Available for Images Created When Someone Was Under 18
People who were under 18 when the image was created should use the National Center for Missing & Exploited Children’s Take It Down service instead of StopNCII. StopNCII directs people in this situation to Take It Down.
6. U.S. Victims Can Report Covered Platforms to the FTC
Under the TAKE IT DOWN Act, covered U.S. platforms must provide a process for requesting the removal of qualifying non-consensual intimate imagery. The Federal Trade Commission is responsible for enforcing the platform requirements, and people can report potential violations to the FTC.
7. Australia Provides Image-Based Abuse Reporting
In Australia, the eSafety Commissioner accepts reports involving image-based abuse and can take steps to help remove qualifying content. This includes certain forms of synthetic or AI-generated intimate imagery.
8. Removing One Copy Does Not Remove Every Copy
The main challenge is that deepfake pornography can be copied and modified faster than individual pieces of content can be identified and removed. A practical response is to preserve relevant evidence, record URLs and account details where safe, report the content directly to the platform, and use the appropriate national or specialist reporting service.
Effective removal therefore often requires more than one takedown request. Content may need to be reported across multiple platforms, while hash-based tools can help participating services identify matching copies and prevent some repeat uploads.
FAQs
1. How common is deepfake porn?
Deepfake pornography makes up a very large share of identified deepfake videos. A 2023 study found that 98% of 95,820 identified deepfake videos were pornographic. In 2019, Deeptrace found that 96% of identified deepfake videos were pornographic.
2. How many deepfake porn videos exist?
There is no reliable global count of deepfake porn videos. Researchers identified 14,678 deepfake videos in 2019 and 95,820 in 2023, but these figures came from different studies and do not represent every video available online.
3. What percentage of deepfake porn victims are women?
Women make up almost all identified deepfake pornography targets. Security Hero found that 99% of people targeted in its 2023 dataset were women, while Deeptrace found that 100% of subjects on the deepfake pornography sites it studied in 2019 were female.
4. Are only celebrities targeted by deepfake porn?
No, celebrities and other public figures are common targets, but private individuals can also be targeted. A study of more than 16,000 people across 10 countries found that 2.2% had experienced non-consensual synthetic intimate imagery.
5. Is deepfake porn illegal?
The law depends on the country and the circumstances. The U.S., UK, EU, and Australia have introduced or strengthened laws covering non-consensual intimate imagery and AI-generated sexual content. Local laws may also provide additional criminal or civil remedies.
6. Can deepfake porn be removed from the internet?
It can often be removed from individual websites and platforms, but complete removal is difficult. Copies can be reposted or edited and shared elsewhere. Tools such as StopNCII can help participating platforms identify matching content using digital fingerprints.
Wrapping Up
Deepfake pornography has grown from a niche form of manipulated media into a wider online safety and privacy concern. Research shows a sharp increase in identified deepfake videos, with pornography making up a large share of the datasets studied and women accounting for most identified targets.
Although celebrities and entertainment figures remain highly visible, private individuals can also be affected. The problem is made more difficult by the rapid spread of content across websites, forums, social media, and private channels, as well as the ability to create altered copies that can avoid simple detection systems.
Governments are responding with stronger laws, platform takedown requirements, and greater oversight of AI tools. However, removing one copy does not always remove every version. Going forward, better detection technology, faster reporting systems, stronger platform cooperation, and clearer legal protections will be important for reducing the spread and impact of non-consensual deepfake pornography.
