BetterTech

The episode covers a reasonable breadth of AI governance topics - shadow AI detection, DSPM, PETs, red teaming - but density is low due to extensive host restatements, social padding, and introductory-level explanations that a working practitioner would already know. Actionable, non-obvious claims are scattered and rarely developed beyond a surface level.

“with those new tools you can really get all the info like what's going on in your network. You can uh, integrate with Cscaler or with Palo Alto Networks, you can have your APIs tracked, you can have uh, a browser extension”

“differential privacy which is kind of the golden standard of data anonymization these days. While still hard to implement. NIST has just published the final version or some months ago of its guidance”

The clearest original moment is the guest's distinction between ethics (should we?) versus responsible/trustworthy AI (are we compliant?), which is a genuinely useful conceptual separation. Everything else - shift left, privacy by design applied to AI, three lines of defence - is recycled governance vocabulary with no contrarian edge.

“Ethics. Should we replace unit ABC with um, AI agents and just fire those 40 people? It's legal. But should we do it? So it's more about this should we and not are we allowed to”

“AI governance can be used for AI strategy, right? So it's not, not only about punishing people. Oh my God, it's terrible, right? But hey, what are they using it for?”

Katharina Corner is a credible AI governance practitioner with a legal background and hands-on consulting experience, and she demonstrates real familiarity with the regulatory and technical landscape. However, she leads a six-person team at a mid-sized IT solutions firm and her frame of reference is consultative rather than that of an operator who has built and scaled governance programs from inside a major enterprise.

“I'm the principal consultant in the AI Governance, AI Risk and AI Security team. We are about six people so far, uh, and growing”

“I mean when I was uh, thinking about this question ahead of time, I am sorry but really what came to mind was Microsoft”

The episode references a reasonable number of named frameworks, standards, and tools (NIST AI RMF, ISO IEC 42001, OWASP LLM Top 10, Colorado AI Act, MIT AI Risk Repository, Cscaler, Palo Alto Networks) and one concrete real-world failure case. However, there are no dollar figures, adoption metrics, project timelines, or detailed client case studies; the Microsoft example is generic and the Dutch government case is named but not elaborated.

“the entire uh, um, Dutch, um, uh, government had to step back three years ago because the AI system they were using for social uh, um, support, Social Security support was just completely discriminative”

“MIT AI Risk Repository, which is tracking over a thousand AI risks”

The host is knowledgeable and keeps the conversation substantive, but she frequently restates the guest's answers at length, asks leading or compound questions, and rarely pushes back or challenges a claim; the episode reads more like a friendly catch-up between colleagues than a disciplined interview designed to extract maximum insight.

“Is that about kind of how it's working today or we'd like it to work?”

“Okay. So it's just cracking break, trying to break things and figure out why. Okay, um, that's what I thought it was but I just wanted to double check.”