Artificial Intelligence Task Force Workstreams

AI is an important global topic that is getting increased public attention, including from key Supranational organizations. Modern AI approaches link algorithms comprehensible to relatively few experts (including actuaries working in the practice area of data science) with almost all aspects of professional and everyday human activities. Actuaries, with their unique expertise and skill set, have a role to play in identifying and managing the risks involved in designing and using AI and also in exploiting responsibly the opportunities AI offers. As the worldwide organization of actuarial associations, the IAA can play a role in facilitating the inclusion of actuaries in the AI conversations at a global level.

Phase 2 - 2025/26

The second phase of work in 2025/26 will focus on how the IAA can support actuaries around the world to become fully AI enabled. In time, it is inevitable that AI enabled actuaries will replace actuaries without AI. Four workstreams will work on the deliverables in the second phase.

Workstream 1: Engagement and Foundations (led by Tim Bishop)

» Members

This workstream focuses on generating interest in AI within the actuarial community as well as helping actuaries get started in the space. The goals include promoting engagement, building and growing the community of actuaries using AI, as well as providing an easy way for actuaries to get started in AI.

Workstream 2: Research and Advancement (led by Henning Wergen and Valerie du Preez)

» Members

This workstream focuses on organizing and surfacing intermediate knowledge as well as encouraging the advancement of the AI-enabled actuary through practical research. The goal of this workstream is to meet the AITF objective of scanning the AI environment relevant to actuaries and supporting the education of actuaries in this field.
This workstream should coordinate with the Engagement and Foundations Workstream to help actuaries with a joint goal of engaging and educating.

Workstream 3: Case Studies and Tools (led by Simon Hatzesberger)

» Members

This workstream organizes case studies and toolkits which are used in AI across all topics, including machine learning, deep learning, NLP, agentic AI, fairness testing. The goal of this workstream is to collect, house, and organize tools and workbenches for actuaries to use in practice.
This workstream should coordinate with the Research and Advancement Workstream in the event there are assets which contain both knowledge and tools, e.g. papers linked to practical packages or repositories. These assets can be housed in the outputs of both workstreams. Also, to the extent that the Research and Advancement Workstream has a taxonomy, this workstream could use that to organize the how-to guides.
This workstream should coordinate with the Engagement and Foundations Workstream on any how-to guides which may have practical use (e.g. getting started in Python or incorporating open AI into Python). These assets can be housed in the outputs of both workstreams. Indeed, for basic how-to guides, the Engagement and Foundations Workstream could use Case Studies and Tools to drive engagement.

Workstream 4: Adoption Framework (led by Bogdan Tautan)

» Members

This workstream provides a framework to help actuaries put it all together and evaluate adoption strategies. In particular, the framework should include considerations such as best practices, professionalism, governance, ethics, as well as practicality.