The SOA and the Age of AI
Exploring potential impacts on professions July 2024Photo: Shutterstock/ImageFlow
Editor’s note: Society of Actuaries (SOA) CEO Greg Heidrich originally presented some of this content about AI tools and research during the 2024 ReFocus Conference, which was sponsored by the SOA and the American Council of Life Insurers (ACLI).
The challenge of adapting, building, and sustaining professional workforces in the “Age of AI” is one every actuary and actuarial employer is facing either in their own work or as employers of highly skilled and experienced professionals.
Many have discussed the rise of artificial intelligence (AI) and big data in the insurance industry and the issues they pose. These tools are transforming the industry; there’s a major trust issue around their proper use; industry leaders have a responsibility to build an ethical framework for their use; and training in the responsible use of these tools is critical.
But over the past 18 months, we’ve also seen an explosion of interest in AI tools. Leaders of the profession and industries in which actuaries work must build and implement effective AI strategies—although many are still struggling to understand the technology well enough to do so.
In some respects, these demands are little different from the pressure leaders faced in the 1990s to build an “internet strategy”; in the 1950s to develop “computer strategies”; and probably in the 1880s to build “electrification strategies.”
AI is a general-purpose technology that ultimately will be used in all industries and across many areas of human life for a currently unimaginable array of uses and benefits.
Impact on Work
The AI revolution will have a wide range of impacts on the nature of work itself. This will apply to the actuarial profession the SOA serves as well as to the work and careers of employees across the insurance industry, where most actuaries work. The industry runs on the brains and skills of the people it employs. Increasingly, it also will run on the brains and skills of AI with a strong complement of human direction and leadership. But the mix of skills needed, the distribution of tasks done by each, and the ways the necessary human skills will be developed are far from clear.
In the SOA’s headquarters office, we have an arithmeter, a cylindrical mechanical slide rule Elizur Wright patented in 1869. At the time, it was the apex of calculating technology, and actuaries throughout the industry used it. While we treasure and display it proudly, it’s a useless relic … good only for museum displays and, of course, there’s no one who knows how to use it.
The same likely will be true someday of the tools actuaries and most other insurance professionals use today. And, if we’re to believe some of the things being written about AI’s possible impact on human work, the same may be said in the future about whole occupations or professions.
It’s the job of a professional society like the SOA to think about the future of the profession and what we do. But it’s also our job to ensure that actuaries are equipped with the knowledge and skills necessary to ensure their success and the success of their employers far into the future.
There are a variety of issues AI will create for professional work, including actuarial work, but first, some context is needed.
Figure 1 lists competencies many professionals, including actuaries, must have. The list is drawn from research the SOA has done to ensure our credentials deliver the skills employers need from actuaries. The highlighted skills overlap some of the most interesting, exciting, (and perhaps a little unsettling) capabilities now being developed most quickly in advanced AI tools.
Figure 1: AI Will Change What Professionals Do
| Category | Skill |
| Higher cognitive skills |
|
| Social and emotional skills |
|
| Technological skills |
|
Key: Skill sets overlapping opportunities from AI
In 2019, Richard and Daniel Susskind of Oxford University published their seminal work, The Future of the Professions: How Technology Will Transform the Work of Human Experts. They concluded that we are “on the brink of fundamental and irreversible change in the way the expertise of specialists is made available to society” and “technology will be the main driver.”
Early this year, a team of labor economists at the International Monetary Fund (IMF) released an important new paper titled “GenAI: Artificial Intelligence and the Future of Work.” They are among the many labor economists trying to understand how AI is likely to affect human work and workers’ lives. The IMF team noted that in advanced economies, about 60% of all jobs currently are exposed to AI, especially those heavy on cognitive task-oriented roles. Of these, they estimate that “about half will be negatively affected, while the rest could benefit from enhanced productivity through AI integration.” The IMF team believes AI is bringing profound change—more rapidly than most of us realize—to the global economy and the world of work akin to a new industrial revolution.
The IMF analysis provides deeper insights on the economics of AI and human work. AI tools are rapidly being developed with the capability to replace specific human work tasks, creating competition with humans for some tasks and reducing the income workers can command for those tasks. At the same time, other tasks done by AI are best seen as complementary to the work of human professionals. These may be tasks an AI could do better than raw human power, but when combined with human interpretation, context, and direction, they will produce better or more socially trusted results. Rather than displace complementary tasks, AI will make them more valuable, boosting their productivity and increasing their value.
The IMF team believes the coming AI wave will lead to deep and wide productivity gains, boost investment in AI tools, and increase overall labor demand. But the gains likely will be unevenly distributed with significant winners and losers. And because many of the most highly educated workers are best equipped to adjust to the changes, there also could be an increase in income inequality across occupations and countries.
These issues are not escaping public attention. The Pew Research Center has documented the growing rise in public unease with the increased use of AI tools in daily life. There’s a global race underway among governments to find ways to secure the undeniable benefits of AI while addressing the undeniable risks. Over the past year, prominent leaders of the technology community at the forefront of AI development have called for a moratorium on AI development until governments and society can “catch up.” And in January, Mustafa Suleyman, CEO and co-Founder of Inflection AI and DeepMind, published an article in Foreign Affairs magazine calling for a global “AI containment strategy.” These statements speak volumes about how unprepared some of the AI field’s leaders believe the world currently is for AI.
The SOA’s Role
I believe the SOA has three main responsibilities in response to this issue:
- Help our members and candidates gain the skills needed to succeed in an AI-driven world
- Embed AI tools productively and safely in our work and services
- Convene, participate, and lead in the consideration and discussion of the issues AI poses for the profession and industry
Preparing Candidates and Members
Helping our members and candidates prepare is our most important obligation and is embedded in our mission. This will take at least three forms:
- Preparing educational programs for working actuaries to help them understand the AI-driven challenges and opportunities they face and begin building the skills they’ll need
- Identifying the skills aspiring actuaries need and building those into our credentialing programs
- Conducting research on AI tools actuaries need and the issues they’ll face in using those tools
In 2023, we offered members nearly 20 different AI-focused educational programs. We’ve completed and updated research intended to introduce actuaries to possible uses and ethical considerations of AI tools, including ways to avoid unfair bias from AI. Members should expect to see many more AI-focused learning and professional development opportunities using the full range of our educational programs and channels.
Bringing AI concepts and skills into the credentialing system isn’t far behind. It will mean identifying the underlying skills tied to successful AI usage by actuaries. We’ve already embedded a significant amount of data science and predictive analytics in the curriculum. Over the next several years, we’ll identify AI-related competencies, test them as teachable skills, and add them to the curriculum if not already there.
We’ll also complete new research aimed at monitoring, understanding, and communicating to members how AI tools are moving into actuarial work and how actuaries can best complement, enhance, and incorporate those tools in their work.
Embedding AI in SOA Operations
We’re implementing policies for responsible use of AI tools by staff and volunteers as well as for the disclosure of AI usage to users of our services. We’re creating greater website personalization, improving website search functions, and providing AI-generated personalized communication.
Looking to the future, we expect to explore ideas such as using AI to support candidate learning by creating individualized exam preparation support, providing feedback to candidates, or, perhaps preparing early drafts of exam questions or providing assistance in first-draft exam grading. We’ll need to create clear guidance for candidates on when and how they’re allowed or expected to use AI tools in their learning and exams. We also can anticipate bringing AI much more directly into our benchmarking research on insurer experience, mortality, or health morbidity patterns, possibly identifying and explaining previously unperceived patterns and conclusions.
Convene, Participate, and Lead
The SOA is the world’s largest actuarial credentialing and professional body. Our work has a major impact on what it means to be an actuary and the work of the profession itself. We have a duty to help create and support opportunities for the profession to explore issues critical to its future and the future of the industries in which we work.
We’re still defining the scope of our work in this space, but I’d like to mention two early initiatives. First, through the efforts of the International Actuarial Association, the global profession has launched explorations of the implications of AI for actuarial professionalism and ethics, actuarial education, AI governance, and AI-focused innovation of actuarial tools and methods. The SOA is supporting this effort and bringing a strong team of professionals to support these workstreams.
Second, the U.S. Department of Commerce recently announced the creation of the U.S. AI Safety Institute Consortium. This consortium brings together AI creators and users, academics, government, industry researchers, and other organizations to establish a new measurement science focused on promoting the safe and trustworthy development and use of AI, particularly in its most advanced forms. The SOA’s Research Institute was selected to be included among 200 leading institutions across many industries and sectors involved in this work. Although the effort is just beginning, we’re committed to supporting it and will work to bring actuarial perspectives into each of the workstreams.
Finally, as we’re developing our programs this year, we expect to offer many opportunities for AI-focused discussion. As we do so, we’ll look for opportunities to partner with other institutions and individuals to contribute to these conversations whenever and wherever they occur.
Final Thoughts on AI and the Actuarial Profession
We’re just beginning to explore and understand the scope and implications of AI for the actuarial profession and the work of actuaries. The SOA is working hard to develop our capabilities and services in AI because we know it’s critical to the future of the profession and the members and candidates we serve.
Statements of fact and opinions expressed herein are those of the individual authors and are not necessarily those of the Society of Actuaries or the respective authors’ employers.
Copyright © 2024 by the Society of Actuaries, Chicago, Illinois.