Why clinical risk, not just accuracy, should guide automation
In Episode #46 of Unboxing AI, Mads Jarner Brevadt, Co-founder & COO of Radiobotics, joins Vidur Mahajan to unpack what it really takes to scale AI in musculoskeletal radiology.
From validation across demographics to the ethics of automation, Mads breaks down how Radiobotics became a global player with RBfracture™ while staying laser-focused on the emergency workflow.
Key discussion points
- It’s not about what AI is good at – it’s about what happens when it fails. Why clinical risk, not just accuracy, should guide automation.
- We trained one model for all anatomies and all ages – from toddlers to the elderly. How Radiobotics avoided model fragmentation.
- You don’t need 50 AI tools. You need one that works – everywhere. Why generalisability needs more than regulatory clearance.
- Diagnostic accuracy is just the start – workflow integration is where value shows up. Lessons from emergency deployments.
- We’re not just building AI, we’re building trust. How Radiobotics trained, deployed, and validated globally – without local teams.
