AI Forensics

AI Forensics is a European non-profit that investigates influential and opaque algorithms and holds major technology platforms accountable. The organization was founded in 2021 by Claudio Agosti, Marc Faddoul, and Salvatore Romano, building on previous platform accountability work. Claudio Agosti had started Tracking Exposed in 2016 with the Italian hacktivist community, while Marc Faddoul conducted research on YouTube's recommender system at UC Berkeley. The full transition from Tracking Exposed to AI Forensics was completed in May 2023. The organization conducts independent, high-profile technical investigations to uncover and expose harms caused by platform algorithms. Their team includes experts from computer science, law, ethics, sociology, psychology, and communication. They develop adversarial auditing tools using methods such as sock-puppeting, scraping, and user data donation, which allow them to collect behavioral data independently from platforms and guarantee data integrity. These tools are released as free software to democratize algorithmic accountability research. AI Forensics has conducted numerous impactful investigations. Their TikTok audit, conducted with the Algorithmic Transparency Initiative and Amnesty International, revealed how TikTok's algorithms could lead young users to harmful content. Their Meta investigation exposed a major moderation loophole enabling pro-Russian propaganda ahead of EU elections, leading the European Commission to launch investigation proceedings. Their Microsoft Copilot research found about a third of election-related answers contained factual errors. In January 2026, they published a major investigation into Grok generating sexualized images of women and minors, analyzing over 20,000 AI-generated images and 50,000 user requests. The organization regularly advises EU regulators, including the EU Commission, the French Digital Council, the French media regulator ARCOM, and has provided testimony to the French Senate and European Parliament. Marc Faddoul, co-founder and Director, is recognized as a leading voice on algorithmic accountability and has been quoted in major publications including The Wall Street Journal, The Guardian, and Le Monde. AI Forensics is registered as AI Forensics Limited in the United Kingdom, incorporated on March 10, 2023. The organization is supported by funders including the Mozilla Technology Fund, the European AI & Society Fund, Luminate, NGI Search (EU Horizon), Civitates, and Open Society Foundations.