3280: Yobi and the Future of Ethical AI at Scale

What if companies could tap into powerful behavioral AI without compromising user privacy or crossing legal lines? In this episode of Tech Talks Daily, I sit down with Frank Portman, CTO of Yobi, to explore how his team is building foundation models grounded in real user behavior, backed by ethically sourced and consented data.

Frank shares how Yobi is taking a distinct approach. They're not building large language models or racing to dominate generative AI headlines. Instead, they're focused on data integrity, transparency, and security from day one. With a strategic partnership with Microsoft Azure, Yobi delivers models that run directly within a customer’s environment. That means privacy is preserved, data stays protected, and companies still benefit from intelligent, adaptive systems.

We unpack how Yobi avoids risky use cases like financial underwriting and healthcare, how their models are trained to avoid demographic bias, and why they actively reward systems for being bad at guessing personal traits. This isn’t just about compliance. It’s about designing products that work better because they’re built responsibly.

Frank also opens up about Yobi’s internal culture, his belief in first-principles thinking, and how empowering engineers to "place bets" drives innovation. He offers insight into what the AI industry must learn quickly from recent missteps, including data misuse and growing public skepticism.

If you're exploring AI solutions and wondering how to build or buy systems that scale without cutting ethical corners, this conversation delivers clarity, honesty, and direction. Are you ready to rethink what responsible AI should look like inside your company?