A new study has raised questions about how major social media platforms, including YouTube and X (formerly Twitter), may be helping users discover AI-powered deepfake image generation tools despite having policies that restrict sexually explicit content.

The research, conducted by the Institute for Strategic Dialogue (ISD), suggests that millions of users were referred from social media platforms to websites offering AI tools capable of generating non-consensual or explicit synthetic images. The findings have renewed concerns over online safety, platform accountability, and the effectiveness of existing content moderation systems.

What the Study Found

According to the ISD report, researchers analyzed 10 of the most widely used AI deepfake image-generation websites and applications to understand how users were discovering these services.

The study examined referral traffic between December 2025 and March 2026 and found that social media platforms collectively generated more than 5.7 million visits to websites providing AI-powered explicit image creation tools.

Researchers said the findings highlight the significant role that recommendation systems and searchable content may play in directing users toward such services.

YouTube and X Accounted for the Largest Share

Among the platforms analyzed, YouTube was identified as the largest source of referral traffic.

According to the report:

• YouTube generated approximately 1.82 million visits, accounting for more than 30% of the total referrals.

• X (formerly Twitter) ranked second, contributing over 1.3 million visits to the analyzed websites.

The report notes that these figures represent referral traffic and do not necessarily indicate that the platforms officially endorse or promote such services.

Search Results Allegedly Led Users to AI Tools

Researchers observed that users searching for terms related to AI image manipulation were reportedly shown videos reviewing or demonstrating these tools.

Some videos allegedly included:

• Direct links to AI image-generation websites.

• Promotional codes offering free credits.

• Tutorials explaining how to use the platforms.

• Reviews comparing different AI-generated image tools.

The researchers argue that this type of content may make it easier for users to discover services capable of generating explicit synthetic images.

Researchers Raise Policy Questions

The ISD report argues that the findings raise concerns about whether platform moderation systems are effectively enforcing existing content policies.

According to the researchers, while many platforms prohibit explicit or pornographic content, similar standards should also apply to content that promotes or directs users to AI tools capable of creating such material.

They suggest that recommendation systems, search visibility, and referral links should receive greater scrutiny as AI-generated content becomes increasingly sophisticated.

Why the Issue Matters

The rapid growth of generative AI has intensified concerns over the misuse of deepfake technology.

Experts have repeatedly warned that AI-generated explicit images can contribute to:

• Online harassment.

• Identity misuse.

• Non-consensual image creation.

• Privacy violations.

• Digital impersonation.

Governments, regulators, technology companies, and civil society organizations around the world are continuing to explore ways to balance technological innovation with stronger user protections.

Platform Policies and Ongoing Debate

Both YouTube and X have community guidelines governing harmful, explicit, and abusive content. However, debates continue over how these rules should apply to videos, links, tutorials, or external websites that may facilitate access to AI-generated explicit content.

The study adds to the broader discussion about whether platforms should take additional measures to limit discoverability of services that could be used to create harmful synthetic media.

As AI technology evolves rapidly, experts expect greater attention from regulators and technology companies on strengthening safeguards against misuse while maintaining open access to legitimate AI applications.

Disclaimer: The findings summarized in this article are based on a report published by the Institute for Strategic Dialogue (ISD). The reported referral traffic does not, by itself, establish that any platform intentionally promotes or endorses the external services mentioned.