AI-generated content in Europe 2026: adoption and detectability

Since 2 August 2026, Article 50 of the EU AI Act requires AI-generated content to be marked and disclosed. This data report asks how much synthetic content is now produced in Europe, and how reliably it can still be recognised once it circulates. It combines Eurostat's 2025 surveys of individuals (isoc_ai_iaiu) and enterprises (isoc_eb_ai), the French Baromètre du numérique 2025 (Arcep, Arcom, CGE, ANCT) and STATEC enterprise data for Luxembourg with detection benchmarks for text, images and video, including an evaluation on RA-Bench, a public set of AI videos generated from real crisis footage. One in three people aged 16 to 74 in the EU used generative AI tools in 2025, one in ten enterprises used AI to generate pictures, video or audio, and a third of people in France have used AI to create images, photos or video, while only one in five users always checks the output. Detection of original files is reliable for text, images and video; after resharing, detection of AI video drops sharply while real footage is rarely flagged. Files: the report in English and French (PDF) and all tables as CSV (CC BY 4.0). Figures derived from the French survey are published separately under ODbL on the report page. Online version in English, French, German and Spanish: https://www.wasitaigenerated.com/research/ai-generated-content-europe-2026-report. Conflict of interest: the author's company develops the detectors evaluated in the report.