A recent study alleges that YouTube and X referred millions of users to websites offering AI-powered tools capable of generating non-consensual sexual deepfake images, despite platform policies restricting harmful content.

As artificial intelligence becomes more accessible, concerns over its misuse continue to grow. A new report claims that major social media platforms, including YouTube and X (formerly Twitter), are inadvertently helping users discover third-party AI tools capable of creating non-consensual sexual deepfake images.

The findings have renewed debate over whether existing content moderation systems are sufficient to prevent the promotion of harmful AI-powered services.

Study Tracks Traffic to AI Deepfake Websites

According to the report, researchers analysed 10 of the most popular AI image-generation websites and apps that allegedly offer tools for creating sexually explicit deepfake images.

The study found that between December 2025 and March 2026, social media platforms collectively directed more than 5.7 million visits to these websites.

YouTube and X Reportedly Led Referrals

The report states that:

• YouTube accounted for approximately 1.82 million referrals, representing more than 30% of the total traffic analysed.
• X (formerly Twitter) reportedly generated over 1.3 million referrals to the same category of websites.

The figures are based on the researchers' analysis and have been cited as evidence of gaps in online content moderation.

How Users Were Reaching These Websites

According to the study, users searching for phrases such as "undress app" were shown videos demonstrating or reviewing AI-powered image-generation tools.

Researchers claim that some of these videos allegedly included:

• Direct links to third-party websites.
• Promotional or referral codes.
• Free credit offers encouraging users to try the services.

The report argues that such content made it easier for users to access platforms capable of generating harmful or non-consensual AI images.

Researchers Raise Concerns Over Platform Policies

The researchers behind the report argue that the findings expose potential shortcomings in content moderation.

While YouTube and X have policies prohibiting certain forms of harmful or explicit content, the report suggests that promotional material directing users to third-party AI services may still remain accessible in some cases.

The study recommends that moderation efforts should focus not only on explicit content hosted directly on social media but also on content that promotes services allegedly enabling the creation of abusive AI-generated imagery.

No Official Response Confirming the Findings

At the time of writing, the report's conclusions remain the claims of the researchers. Any responses or actions from YouTube, X, or the operators of the websites mentioned would depend on their official statements and internal policy reviews.

Both platforms maintain community guidelines aimed at limiting harmful content and regularly update their moderation systems to address emerging forms of online abuse.

Growing Focus on AI Safety

The rapid rise of generative AI has created new opportunities for creativity and productivity, but it has also increased concerns about misuse, particularly involving deepfake technology.

Governments, technology companies, and digital safety experts continue to explore stronger safeguards to reduce the spread of non-consensual AI-generated content while balancing innovation and freedom of expression.

Disclaimer: This article is based on the findings of a published research report. The referral figures and allegations cited are those of the researchers and should not be interpreted as proof of intentional promotion by any platform. Official responses or further investigations may provide additional context.