the physical side of AI readiness - Automation-Readiness

---

The clip divides AI risk into technical risks intrinsic to models, including hallucinations, discrimination and generative AI intellectual property infringement, and non-technical risks amplified by reliance on AI, such as business interruption when systems fail.

• The speakers argue that technical AI risks will require specialized underwriting and dedicated products, and that non-technical risks are more likely to be addressed through existing coverage structures with endorsements or affirmative language. That is their view of where the market is heading, not a description of what any policy does today.

Insurance Journal Clips · 9 minute YouTube video

Simulation and digital twin testing let a robot fleet or a process change be exercised in a physics accurate model before it runs on a floor. This is NVIDIA describing its own platform, so the performance claims in it are the vendor's rather than ours.

• What it raises for a risk conversation: simulation runs, software in the loop testing and continuous sensor feedback are records. Whether an insured can produce them, and whether they were created before the equipment was introduced rather than after, is a change control question this page asks directly.

NVIDIA · 3 minute YouTube video