Episode Details

AI is becoming increasingly capable—but the harder question may not be what AI can do. It may be what we should allow it to do.

In this episode of Thought Atlas, I sit down with Latha Karthigaa, an AI governance expert and technology leader who serves as Head of AI Governance at the Global AI Certification Council. Latha began her career as a software engineer, completed a PhD in software engineering, worked for a software company, and later started a digital marketing company, which she says she has been running for around eight years.

Her work in AI governance has led her to focus on a question that sits at the heart of responsible technological development: how do we make sure increasingly powerful AI systems remain useful, trustworthy, and safe for people?

Latha makes an important distinction between compliance and governance. Compliance, in her view, is about following rules and regulations. Governance goes a step further by asking whether an AI system should exist in the first place, and whether it could negatively affect individuals, organizations, or the wider ecosystem.

We explore why she believes “governance first, technology second,” and why human oversight remains essential when AI systems make consequential decisions. Latha discusses several AI incidents that she says demonstrate the potential consequences of poorly governed systems, including automated decision-making in Dutch childcare benefits and an AI hiring system that reportedly discriminated against women.

The conversation also examines trust. Can we trust an AI system in the same way we trust another human? What happens when people become emotionally attached to chatbots? And how should we think about AI systems that can influence vulnerable people?

Latha argues that responsible AI requires more than technical expertise. For her, empathy is non-negotiable. She believes future AI governance professionals need awareness, open-mindedness, a commitment to building things that genuinely benefit people, and an understanding of the failures that have already occurred.

We also discuss explainability: if an AI system makes a decision that affects someone’s healthcare, finances, education, employment, or public services, should it be able to explain how that decision was reached? Latha argues that explainability and human review are critical considerations for high-risk AI systems.

Then we turn to a more personal and philosophical question: What is AI doing to the way we think?

Latha believes AI can dramatically improve productivity and automate tasks, but she is concerned that people may begin relying on it even for simple thinking and problem-solving. She argues that younger people in particular should learn to think independently before becoming heavily dependent on AI tools.

Finally, Latha looks ahead at the next decade and shares her personal perspective on the possible convergence of AI, artificial general intelligence, quantum computing, and robotics—and what that could mean for work, jobs, and everyday life.

Her central message throughout the conversation is simple: AI governance is ultimately about humanity.

Latha shares her daily AI-incident content on her YouTube channel, Latha - AI - Governance, and says listeners can also connect with her on LinkedIn by searching for Latha Karthigaa. She is also writing a book about AI governance and the incidents she has studied, which she says she expects to release around the middle of next year.

Connect with Latha through her YouTube channel or LinkedIn, as mentioned above.

Listen to Thought Atlas: joinpodmatch.com/thoughtatlas

Food for thought: “An idiot with a plan can beat a genius without a plan.”