Deepfake report finds US & X lead global incidents

identifAI has published a report analysing more than 10,000 deepfake incidents worldwide. It found that the United States accounted for 46.9% of recorded incidents.

Covering incidents identified between January 2020 and March 2026, the report points to social media as the main distribution route. X accounted for 51.2% of incident propagation, ahead of TikTok at 21.1%, YouTube at 10.0% and Facebook at 8.2%.

Political manipulation made up 24.6% of the cases examined, making it one of the largest categories in the dataset. Fraud-related incidents accounted for 20.1% of the total. The report also says synthetic media has been used to shape geopolitical narratives involving political figures and state entities.

After the US, the UK was the second most targeted country, with 8.2% of global incidents. India accounted for 7.2% and Israel 6.6%, followed by Iran at 2.9%, South Korea at 2.1% and Australia at 1.8%.

Attack Patterns

Video was the most common format among the incidents analysed, representing 45.6% of cases. Mixed media accounted for 25.2%, images 17.4% and voice cloning 10.5%.

The report was compiled from public indicators rather than proprietary telemetry. It included only incidents in which a deepfake had been deployed in a real-world context and was supported by English-language media reporting or explicit mentions of synthetic media activity on social media platforms.

According to the methodology, the underlying data was processed through a large language model-based classification system that extracted entities, identified formats and grouped incidents into categories. identifAI then applied percentage-based statistical modelling to reduce distortions that can arise from raw volume counts.

The findings suggest synthetic media has moved beyond isolated hoaxes into broader use across politics and fraud. The report argues that such material is increasingly designed not only to mislead people but also to pass verification checks and spread quickly through recommendation systems.

Platform Role

X was identified as the largest propagation channel in the incidents studied. TikTok was the second most used platform, while Instagram, Telegram and WhatsApp were each linked to smaller shares.

The report links the spread of deepfakes to platform systems that amplify highly engaging posts. It also calls for liability frameworks requiring social media platforms to take greater responsibility for the algorithmic distribution of synthetic media.

Other recommendations include wider use of digital provenance standards and technical controls aimed at identity and media verification. Measures cited include protocol-level camera injection defences and multi-modal biometric checks designed to compare vocal and facial analysis.

identifAI is an Italian startup focused on detecting AI-generated images and videos. It said it raised €5 million in a funding round led by United Ventures in July 2025.

Dr Marco Ramilli, chief executive officer and co-founder of identifAI, said: “Our findings confirm that deepfakes and other synthetic media are now commoditized tools for large-scale extortion and disinformation. As threat actors move toward real-time generation and automated deployment, the window for human intervention is collapsing. Institutional resilience no longer depends on reactive moderation, but on integrating mathematically rigorous, multi-modal detection at the architectural level to uphold truth and defend the integrity of our digital world and institutions.”