Why "Detection" Will Never Solve AI Fraud in Insurance | Jeppe Nørregaard, InReality

Detection is an arms race you cannot win. That's the core argument InReality Founder & CEO Jeppe Nørregaard makes in Episode 167 of InsurTechTalk, and it comes from someone who spent seven years working on AI-based misinformation detection before concluding the entire approach was structurally doomed.His pivot: stop trying to catch fakes after the fact. Instead, cryptographically prove content is real from the moment it's captured, so fake evidence never enters the pipeline in the first place.In this episode:▸ Why Jeppe abandoned AI detection research after watching AI get better at creating misinformation than detecting it▸ What post-quantum cryptography actually means, and why old systems, not new ones, are the real quantum risk▸ Why AI has "absolutely no chance" of breaking a cryptographic proof▸ How Check, Sign, Verify works at the level of an actual smartphone camera▸ The insurance fraud numbers driving urgency: a 71% rise reported by Admiral in 2025▸ Why "undetectable AI" makes traditional fraud detection a losing long-term bet▸ GDPR, the EU AI Act, and why permanent blockchain records are actually a privacy liability▸ Why Jeppe wants fraud numbers to keep rising, not fall, and what a falling number would actually meanLearn more: https://inreality.ioConnect with Jeppe: https://www.linkedin.com/in/jeppe-nørregaard/InsurTechTalk by InsurTech.me is the podcast for insurance professionals, InsurTech founders, and anyone who wants a no-BS look at where the industry is headed. New episodes every week.