"SENTINEL: Securing Age Estimation and Digital IDs against Presentation and Injection Threats."

• Unité de recherche: INNOSUISSE
• Numéro de projet: 108.708 INT-ICT
• Titre du projet anglais: "SENTINEL: Securing Age Estimation and Digital IDs against Presentation and Injection Threats."

Mots-clé (Anglais)

ICT, Artificial/local/extended/networked intelligence, Machine learning

Résumé des résultats (Abstract) (Anglais)

We propose a defensive innovation that involves tech and processes to identify & prevent AI-generated attacks on age estimation (AE) systems mandated globally by regulation and currently deployed widely to protect children from exposure to goods, content or services that may cause them harm.