Risky Science Podcast · Risk Market News

AI Risk, Markets and Modeling the Unknown With Daniel Reti

·42 min·1 clip

Insurance companies are pricing AI risk like it doesn't exist

Chris Westphal interviews Daniel Reddy of Exona Labs about AI risk quantification and insurance. Daniel explains how he and his co-founder, both former Bank of America quants specializing in counterparty credit risk, are applying quantitative discipline to AI risk—a domain with almost no loss history but rapidly growing incidents (50% YoY growth). They discuss how most AI risks fall into two categories: AI-accelerated existing risks (like data breaches) and novel risks without historical precedent. Rather than relying on loss history, Exona focuses on governance and engineering signals: Does the company have an AI security lead? Do they fine-tune models with proprietary data? Are they using uncontrolled API integrations? The conversation explores why catastrophe bonds might be more effective AI safety governance than regulation, and how insurance requirements could incentivize safer frontier lab practices.

As heard by us

A measured look at pricing AI risk before the data catches up.

Risky Science treats AI as an underwriting problem, with Chris Westphal and Daniel Reddy, co-founder of Exona Labs, weighing how to price a fast-moving technology when the loss history is thin and the data trail is still forming.

Read the full review in PlayNext →

Why you'd press play

If AI risk sounds abstract, this one makes it feel like an underwriting problem.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Risky Science Podcast