August 2018

Artificial Intelligence for Actuaries

By Sarah Abigail

In March 2018 the Entrepreneurial & Innovation Section co-hosted a webcast on Artificial Intelligence (AI) with the Actuary of the Future Section. I was honored to moderate the webcast with three company founders who shared their views on the development of artificial intelligence and how it relates to the actuarial profession.

Currently, computer chip manufacturers are working on quantum chips that will possess more computing power than all of the processing chips that exist in the world today combined. We are on the precipice of world altering technology that will scale artificial intelligence platforms to phenomenal levels of power and speed. We can expect AI to be in every part of our lives very soon.

The “AI for Actuaries” webcast was designed to answer some of the pressing questions that actuaries have been asking about AI. Is AI useful in actuarial work? What will the impact of AI be on the actuarial industry and on actuaries themselves? How can actuaries get involved in AI?

Nicholas Yeo Chee Lek, FSA, founder of Nicholas Actuarial Solutions, discussed the current challenges that are being faced with the use of IBNR robots in actuarial applications. Currently, very well-organized data needs to be inputted for the robot to provide pragmatic actuarial solutions. The dream is to have a fully automated AI system that can organize, calculate and judge large actuarial data sets. Nicholas shared that with improved computing power, such a futuristic actuarial system is possible. However, only a handful of actuaries are working on it.

Shankar Vaidyanathan, founder of Noonum, shared that highly specialized AI platforms can pull in millions of pages of financial reports and find patterns to help with financial analysis. Shankar presented the basics of AI and suggested that today’s mainstream AI processes have the capacity to learn at the rate of a small toddler. Currently AI can recognize shapes and signs, but it will take time for the AI to do more and take on the tasks that we dream of. Shankar shared that AI certainly will get smarter in time, and its applications will grow exponentially.

Ari Ramdial, founder of Rhodium, shared that today’s AI systems have biases, based on what the programmer feels is important to them. Different actuarial AI platforms could have different outputs based on pre-conceived biases of each programmer. Ari foresees a world where AI systems will not be black boxes, but rather glass boxes where users can understand how the data is being organized, calculated and judged through an AI process. The glass box process is also known as “Explainable AI” or XAI. In the future XAI systems will be simple enough to adopt and implement across all capacities of the actuarial profession.

All three panelists agreed that actuaries have been fairly absent from the conversation about AI and Machine Learning (ML). They were all excited to share their expertise on AI and were hopeful that the presentation will influence more actuaries to get involved in AI developments. In an effort to open up more conversations about artificial intelligence within the actuarial profession, a hashtag was setup #AIforActuaries to promote the conversation on social networks.

Will the future of the actuarial profession be a handful of coders programming AI algorithms? Maybe. What we know for certain is that AI will dramatically change all industries. In the coming decade, all actuaries will use AI and rely on its functions to make faster and more reliable decisions. Actuaries who enter the AI conversation now can significantly influence the shape and direction of AI’s actuarial developments. Those who get involved early will also likely become the ambassadors of the actuarial profession to the tech world.

Sarah Abigail is co-founder and strategy consultant at Ironbound Consulting Group in New York City. Sarah can be reached at