Trustworthy AI
with Yiannis Kanellopoulos, Hosted by Heather D. Couture | Impact AI by Pixel Scientia
22/04/2023
3 MIN READ /
In this episode of the Impact AI podcast, Heather D. Couture speaks with Yiannis Kanellopoulos, Founder and CEO of code4thought, about what it takes to build and evaluate AI systems that are truly trustworthy.
As AI technologies become more embedded in high-stakes domains like healthcare, finance, and HR, the conversation around quality and ethics is no longer a theoretical one. Yiannis shares real-world insights from his experience leading independent audits and testing of AI systems across sectors, highlighting what trustworthiness looks like in practice—and why it’s business-critical.
Topics Covered:
Why trust in AI must be earned, not assumed
What makes AI systems trustworthy, from governance to technical safeguards
Real examples of testing AI models for bias, explainability, and robustness
How context shapes the evaluation of model risk and fairness
The role of independent audits and “last-mile analytics” in AI deployment
How stakeholder needs and transparency affect model adoption and trust
Why trustworthiness is becoming key even for investors during AI due diligence
Strategic advice for AI startups on building trust from the start
Yiannis also unpacks the notion of “last mile analytics”, the final layer of independent, context-sensitive AI validation—developed to catch blind spots that internal development teams may miss before deployment.
“Trustworthiness should not be an afterthought. If you integrate it from the start, it pays off—not just technically, but also from a business perspective.” – Yiannis Kanellopoulos
Explore our Trustworhty AI services and iQ4AI, our AI quality testing platform to learn more about how we help organizations evaluate AI systems across bias, performance, explainability, and risk and achieve both performance and compliance.