AI & Machine Learning

"Slopsquatting" exploits AI coding hallucinations by planting malicious packages with fictitious names that automated tools may inadvertently import into software.

As AI scales, insurers relying on sampling to audit AI decisions may be kidding themselves.

As agentic AI moves from recommending actions to executing them, insurers need an authority layer that defines who an agent acts for, what it may do and where its power must stop.

The key foundation for AI implementation is not the model. It is the data the model relies on.

Underwriting AI systems are designed to assist rather than replace human judgment, but this efficiency trade-off may eliminate how juniors become experts.

Auto and trucking litigation now drives 33% of carrier spending as nuclear verdicts climb, demanding proactive management over reactive response.

Generic AI can summarize and extract data, but vertical AI is needed to provide the context and governance required by high-stakes decisions.

Life insurers carry decades-long risk locked in contracts they can't revise, yet lack systems to continuously test the assumptions beneath them.