Predictive modeling

Digital Insurance spoke with three actuaries about emerging technologies and the challenge of explainability within complex models.

AI-powered catastrophe and risk models use varied sources, including aerial imagery.

Rebuilding artificial intelligence algorithms to explicitly integrate transparency and overcome black box limitations.

Predictive models are a tool to address a business problem, not a solution.

The shortcomings of emerging "climate services" are concerning to some observers.

The predictive model is powered by AI analysis based on data from over 1,500 wildfire losses.

How do you maximize efficiency in the application, quoting, and claims process and not invite new or additional fraudulent activity?

One example of a typical deep learning application is pattern recognition applied to digital images.

Current actuarial techniques are good at quantifying in the aggregate – but less so in the detail.

Research shows rapid adoption of predictive algorithms and digital identity technology will stem the pandemic-driven fraud deluge. Insurance fraud causes more than $80 billion in losses annually in the U.S. alone, according to the report.