What Happens When AI Commits A Crime? Who Goes To Prison?

We've spent years worrying about AI becoming smarter. But what happens when it becomes autonomous? This week we explore the latest AI models that have demonstrated lying, deception and unexpected behaviour during safety testing and ask the question nobody seems to have a straightforward answer to: If your AI commits a crime, is it your fault? Our answer is pretty straightforward though. We also look at Meta's AI accidentally accessing a real company's systems, the growing controversy around AI-powered smart glasses, Elon Musk's plans for AI factories on the Moon, and why AI-generated sexual deepfakes are creating an urgent new legal challenge. Plus, James explains how a new type of clothing could actually make you harder for AI surveillance systems to recognise. Contact us at aihaventaclue.com If you’re looking to scale AI with confidence and real business impact, visit https://kpmg.com/uk/en/services/ai.html to learn how KPMG helps organisations make a difference with AI.