AI flags 242,000 potentially abusive posts ahead of Paris 2024 and Milan‑Cortina 2026
On August 8, the IOC disclosed results from its CAPS system, an AI‑driven tool that scanned millions of online interactions linked to the Paris 2024 and Milan‑Cortina 2026 Olympic projects. The system sifted through 4.3 million posts, detecting roughly 242,000 items that required human scrutiny for possible abuse.
The AI analysis spanned 20,000 accounts and 45 languages for the Paris Games, and 6,000 accounts across 56 languages for the Winter Games. After human review, 24,700 posts were confirmed as abusive – 10,200 originating from the Paris data set and 14,500 from Milan‑Cortina. The investigation also uncovered 18,900 distinct user accounts responsible for the offensive content.
The IOC emphasizes that digital harassment extends beyond virtual spaces, inflicting real‑world damage on athletes’ health, reputation and finances. It urges platforms, organizers and national bodies to adopt comprehensive prevention, education and mitigation measures.