The Last Line of Defence in an AI System Should Never Be the User. Here's Where It Actually Belongs.

And why most enterprise AI deployments have this completely backwards. There's a moment I keep seeing play out across enterprise AI deployments. A system produces an output. The output is wrong — confidently, fluently, plausibly wrong. And the only thing standing between that wrong output and a consequential decision i

This edition was published on LinkedIn. Full text lives there until a first-party archive is available on this site.