Platform for autonomous risk assessment and quantification for cyber insurance policies
A system for autonomous risk assessment and quantification for insurance policies for computer and information technology related risks, including but not limited to losses due to system availability, cloud computing failures, current and past data breaches, and data integrity issues. The system will use a variety of current risk information to assess the likelihood of operational interruption or loss due to both accidental issues and malicious activity. Based on these assessments, the system will be able to autonomously issue policies, adjust premium pricing, process claims, and seek re-insurance opportunities with a minimum of human input.
1. A computer system for autonomous risk assessment configured to execute software instructions stored on nontransitory machine-readable storage media, wherein the software instructions comprise instructions that:
receive near real-time data from a plurality of online data sources, wherein the online data sources relate to a plurality of insured assets, and wherein the data includes data related to one or more of network status, availability of cloud computing platforms, and active or potential cyber attacks;
analyze the data to determine a likelihood of operational interruption or loss from a plurality of computer and information technology related risks from both accidental events and deliberate malicious activity;
perform a plurality of predictive discrete event simulations using results of the analysis;
apply a plurality of predictive algorithms to the data to produce a hazard model and a multi-peril model;
use the hazard model to generate a near real-time risk score for one or more of the insured assets; and
use the multi-peril model to generate a near real-time price for a premium of the one or more insured assets.
2. The computer system of claim 1 , wherein the plurality of predictive discrete event simulations are performed using multiple parallel threads machine learning training.
3. The computer system of claim 1 , wherein the software instructions further comprise instructions that:
train a machine learning model on historical network data, vulnerability data, user actions, risk models, or multi-peril models, or any combination thereof; and
Compare simulation output and real-world outcomes with the trained machine learning model.
4. The computer system of claim 1 , wherein the hazard model and multi-peril model are continually updated with real-time, near-real-time, or regularly polled live data.
5. The computer system of claim 1 , wherein the software instructions further comprise instructions that:
display the hazard model and the multi-peril model for viewing by a user;
permit the user to interact with displayed models;
process user interactions with the models and displays the results;
permit the user to reverse any of their interactions or changes to the models; and
permit the user to save changes or interactions with the models to update the near real-time risk score or the near real-time price for the premium.
6. The computer system of claim 1 , wherein the multi-peril model is a full vulnerability model.
7. A method for autonomous risk assessment and quantification for insurance policies for operational interruption and losses associated with computer and technology related risks, comprising the steps of:
receiving near real-time data from a plurality of online data sources, wherein the online data sources relate to a plurality of insured assets, and wherein the data includes data related to one or more of network status, availability of cloud computing platforms, and active or potential cyber attacks;
analyzing the likelihood of operational interruption or loss from a plurality of computer and information technology related risks from both accidental events and deliberate malicious activity;
performing a plurality of predictive discrete event simulations using the analyzed data;
applying a plurality of predictive algorithms to the data to produce a hazard model and a multi-peril model;
using the hazard model to generate a near real-time risk score for one or more of the insured assets; and
using the multi-peril model to generate a near real-time price for a premium of the one or more insured assets.
8. The method of claim 7 , further comprising wherein the plurality of predictive discrete event simulations are performed using multiple parallel threads.
9. The method of claim 7 , further comprising the steps of:
training a machine learning model on historical network data, vulnerability data, user actions, risk models, or multi-peril models, or any combination thereof; and
comparing simulation output and real-world outcomes with the trained machine learning model.
10. The method of claim 7 , wherein the hazard model and multi-peril model are continually updated with real-time, near-real-time, or regularly polled live data, using an interactive display.
11. The method of claim 7 , further comprising the steps of:
displaying the hazard model and the multi-peril model for viewing by a user;
permitting the user to interact with displayed models;
processing user interactions with the models and displays the results;
permitting the user to reverse any of their interactions or changes to the models; and
permitting the user to save changes or interactions with the models to update the near real-time risk score or the near real-time price for the premium.
12. The method of claim 7 , wherein the multi-peril model is a full vulnerability model.