Ai

Europe’s AI debate increasingly assumes that all providers, architectures, and risks are the same. They are not. As regulators struggle to keep pace with technological change, European consumers risk losing not only access to innovation, but also the right to decide for themselves whom they trust.

Apple’s WWDC 2026 keynote showcased the company’s most ambitious artificial intelligence strategy to date, but many of its most significant innovations remain unavailable to European consumers. As regulators and technology companies continue to clash, the real question is whether European users are increasingly being left behind.

Microsoft’s renewed ARM push is often compared to Apple’s transition to Apple Silicon, but that comparison only tells part of the story. Drawing from first-hand experience with the Windows RT era, this column argues that Microsoft’s latest direction is less about ARM itself and more about preparing Windows for a decade of local AI, hybrid inference, architecture-neutral software, and a new generation of applications designed around intelligence rather than simply connected to it.

This Friday morning I had the opportunity to spend time with final-year business students at UBB and discuss AI, entrepreneurship, Internet of Behaviors, prompt engineering and the growing role of judgment in an AI-assisted world. Somewhere between a one-shot coffee shop demo, a sarcastic AI personality experiment and discussions about the future of work, a larger lesson emerged: AI is not the business. It may become the advantage.

The barrier to building AI products has never been lower. And that’s exactly why most of them will fail. Not because AI doesn’t work-but because building something useful is much harder than generating something impressive.

AI feels unpredictable because it is. Unlike traditional software, it doesn’t execute fixed logic-it generates probabilities. Understanding this changes how you build, debug, and trust AI systems.