The Bridgecast with Scott Kinka

There are genuine moments of insight - particularly Eva Marie's point about regulators grandfathering unaudited models and the mathematical unavoidability of bias - but much of the episode traffics in standard AI governance advice ('don't do AI for AI's sake,' 'start with boring tasks first') that a switched-on B2B operator would have already absorbed. The compilation format and AI-avatar narration bookends cut into actual idea density.

“I recently realized I can exactly, word by word, give the same talks that I gave in 2017 with even the same slides”

“Most of the people there said we slowly grandfather them because we don't want to interfere with the business”

The 'trustworthy vs. responsible' reframe is a genuinely interesting semantic argument, and Eva Marie's blunt assertion that 'you do have to decide on who are you okay with discriminating' is refreshingly direct, but the bulk of advice - map workflows before buying tools, find internal champions, start small - is well-worn change-management doctrine recycled with an AI skin.

“you do have to decide on who are you okay with discriminating”

“no one is going to say, oh, yeah, I'm going to sign off for irresponsible AI, right? That's not a thing”

All three guests are credentialed practitioners rather than pure thought leaders: Reggie Townsend leads a named data ethics practice at SAS with documented policy work; Dr. Muller Stuehler advises governments and global institutions; Elaine Barsoom ran tech innovation at Nike and American Express at genuine enterprise scale. None are superstar-tier but all have done the thing, not just talked about it.

“In the Middle east, for example, fraud detection machines, they are often just bought from the us they don't really work that now because they're not often trained on Middle Eastern information”

“we started really with the business. We mapped where are current state workflows incorporating technology”

The episode has a handful of concrete, named examples - Middle Eastern fraud engines trained on US data, Emirati women over 65 absent from healthcare datasets, Nike's GitHub Copilot rollout with low initial adoption - but hard numbers, dollar figures, adoption percentages, and timeline data are essentially absent, leaving many claims unsupported beyond analogy.

“I can promise you there was no data set of an emirati women over 65 in there”

“credit, um, application model at the bank who gives works perfectly... you'll only discover it five years later when everything fails and nobody is repaying any loans”

The host frames practical, operator-relevant questions ('what does good look like in an AI governance policy for a mid-market business?') but consistently moves to validate rather than probe - there is no meaningful pushback, no numbers challenged, and no productive disagreement across any of the three segments. The compilation format with AI-avatar bridge narration further dilutes conversational depth.

“If somebody says to you, I don't get it, why responsible AI and what's the risk? What's your answer?”

“I love Reggie's medical analogy there. And don't try to cure cancer on day one. Just find where the band aids are needed”