Lloyd’s Launches AI Failure Coverage

Lloyd’s of London, via Y Combinator-backed Armilla, has introduced a new insurance policy that covers legal costs and damages arising from AI system malfunctions (such as chatbots hallucinating false information) marking the emergence of a fresh liability class that actuaries will need to model for frequency, severity, and reserving.

Armilla state they will “assess the AI model, get comfortable with its probability of degradation, and then compensate if the models degrade.” Quantifying the probability of degredation ought to be a fun challenge.

Source: Financial Times

Untapped AI Potential in Actuarial Work

In more news from the Lloyd’s realm: A Lloyd’s Market Association report finds that while 40 percent of London market firms are actively using AI tools, adoption in actuarial and risk functions remains held back by regulatory uncertainty, a lack of in-house AI/ML expertise, and concerns around data quality and model transparency. They see an urgent need for upskilling and governance frameworks if actuaries are to fully harness AI’s quantitative benefits.

There’s a handful of interesting data in their report. Data quality is the highest priority when using AI/ML (weren’t we meant to have that solved a decade ago?) and Chief Actuaries are bigger believers in AI/ML than their CRO colleagues.

Source: LMA

Connecticut Approves AI Roof-Assessment

The Connecticut Insurance Department has green-lit ZestyAI’s aerial-imagery-based roof quality solution for homeowners’ underwriting after a third-party actuarial review, signaling regulatory confidence in AI-driven risk grading and paving the way for P&C actuaries to integrate AI-derived roof condition scores into rating plans.

This isn’t a brand new concept. Aris Detect, amongst others, are developing the same technology in Australia, albeit not chasing the insurance market (yet?).

Source: Zesty

Global Life Insurer Embraces AI Underwriting

Singapore’s Neutrinos has secured a deal with a Fortune 500 life insurer to deploy its AI-driven underwriting platform across eight countries - automating application workflows for “first-time-right” accuracy on mass-market policies and augmenting human underwriters on complex cases - positioning actuarial teams to oversee model calibration, mortality-assumption alignment, and fairness.

No mention about which Fortune 500 life insurer it is, though, and only Neutrino themselves have announced it.

Source: Neutrinos

Redington Weekly brings you the cutting-edge of AI in actuarial science, delivered every Friday. Logo credit Sodacan. © 2025 Redington Weekly. All rights reserved.