L'Intelligenza Artificiale è sicura? Rischi, Regole e Futuro con il Prof. Fabio Roli

Is Artificial Intelligence truly intelligent or just a powerful statistical machine? In this episode of Hello Trust, we host Professor Fabio Roli, Professor at the University of Genoa and principal coordinator of the InfoAIcert project, to explore the frontiers of security and reliability in modern AI. Starting from the famous "cow on the beach" paradox, Professor Roli explains why accuracy does not necessarily equate to robustness and the new vulnerabilities (such as adversarial attacks) that companies must address. In this interview, you will discover: Scaling Law: Why giving AI more data doesn't solve all problems. Trustworthy AI: What "trustworthy AI" means according to European standards. Cybersecurity vs. AI Security: Why traditional software bugs are different from vulnerabilities in learning models. InfoAIcert Project: How to democratize AI Act compliance for small and medium-sized enterprises (SMEs). Competitive Advantage: Why regulation is not just an obligation, but a lever of trust for the market. Video Chapters: 2:36 ChatGPT and the "Scaling Laws" Challenge 5:21 Accuracy vs. Reliability: Why AI Makes Mistakes 8:44 What is Trustworthy AI? 11:00 Adversarial Attacks and Data Security 14:33 InfoAIcert: Certifying AI for SMBs 🔗 Useful Links: InfoAIcert Website: https://infoaicert.ai/