In the previous article, we explored the discussions and initiatives emerging from the IAA AI Summit, highlighting how actuaries worldwide are preparing to navigate the burgeoning wave of artificial intelligence (AI). Building on those insights, this follow-up article explores practical ways that actuaries can actively engage with this technology.
Here are some initiatives and ideas for how we actuaries can learn to harness AI. Some of these sprouted from the IAA AI Summit, others are already-planned initiatives by the CIA. While employers and professional organizations can play a role in facilitating this learning, the responsibility of better understanding AI lies with each one of us.
SPECIALIZED TRAINING PROGRAMS AND WORKSHOPS
Participate in specialized training programs focused on AI and machine learning (ML). Admittedly, there aren’t many AI and ML programs out there tailored for actuaries, but there are plenty of solid foundation programs. Deeplearning.ai has a number of beginner- to intermediate-level courses and specializations, offered in partnership with Stanford University via Coursera, that I have found helpful.
DEVELOPMENT OF AI-POWERED TOOLS
Invest in the development and deployment of AI-powered tools for data analysis, underwriting, claims processing, and customer service. Some companies have invested in building their own foundation model, and others have built, or are in the processing of building, their own internal ChatGPT. The vast majority, however, chooses to hire AI experts to build in-house chatbots or large language models (LLMs) for them.
Encourage actuaries to engage with AI tools through hands-on projects and case studies to better understand their capabilities and limitations. For example, for US$20 a month, anyone can explore the latest ChatGPT 4.0 and Perplexity.ai PRO, and interchange between both depending on the task at hand.
Both ChatGPT and Perplexity have free versions available, but I consider the $20 monthly subscription money well spent.
CIA EDUCATION SYSTEM AND CONTINUING EDUCATION PROGRAMS
Incorporate AI and data science topics into core education and continuing education programs for actuaries. Thankfully, the CIA is already on this! AI content is actively being offered across CIA seminars, its Annual Conference and webcasts, and the topic will be introduced to module content as part of earning the ACIA designation. Ways to incorporate AI in Fellow-level material are also being considered.
Offer courses on programming languages commonly used in AI and data science such as Python and ML platforms like TensorFlow. I’ve found Brilliant to be a neat learning platform for Python, since it incorporates lots of hands-on practice delivered in micro doses. Google offers a host of self-guided courses on TensorFlow. I recommend starting with the free course on Generative AI.
CREATING A KNOWLEDGE-SHARING COMMUNITY
Establish a community or forum where actuaries can share knowledge, experiences and best practices related to AI. Without giving away too much since it hasn’t been publicly announced, I am happy to say that this is in the works – stay tuned for more details.
Organize hack-a-thons, competitions, and collaborative projects to foster innovation and practical applications of AI in actuarial tasks. This is also being worked on by a group of actuarial delegates. Same as above, please stay tuned!
By engaging in one or more of these initiatives, actuaries can develop a deeper understanding of AI, leveraging it effectively in their work, and most importantly, we actuaries can then not only stay relevant, but also thrive amid the wave of AI.
MORE INSIGHTS ON THE IAA AI SUMMIT FROM KEY ACTUARIES
Just like the article featuring Bernice and Tim, I spoke with Matt Bartley, another member of the IAA’s AI Task Force, and answered the questions myself. Together, we share our perspectives on the future of the actuarial profession in the age of AI and provide further insights into the task force members’ AI journeys and experiences.
IAA AITF WORK STREAM: CHANGING ROLE OF ACTUARIES
Q: What got you interested in AI?
A: Both when I was young and throughout university, I had a strong interest in computers, programming and problem solving. AI seems like a natural extension of that. I was also rather inspired by the documentary AlphaGo and how intelligent system design combined with machine learning could produce something so revolutionary.
Q: Why is getting involved in this initiative/topic important to you?
A: As an actuary working in the data science space, I see the opportunity for heightened efficiency and modernization when it comes to using AI in actuarial work. This initiative will allow us to bridge the gap between data science and actuarial work, with an appropriate risk lens.
Q: How do you use AI in your daily work now?
A: I am responsible for Canada Life’s accelerated underwriting program, which is supported by multiple AI models. My daily work in managing the team includes performance monitoring and enhancement of these models. My team is also looking for broader ways to use AI in support of our business unit strategy and from that perspective it feels much more entrepreneurial.
Q: How do you expect to interact with AI in 5-10 years?
A: I think we will start seeing AI embedded in more of our day-to-day devices, but it might take a while for the average consumer to replace existing devices with ones that are more digitally integrated. Perhaps, for example, we will see an AI system that learns over time, based on voice interactions, what everyone’s temperature preferences are in their home and makes automatic adjustments.
I think we’re probably already seeing voice-activated natural language inquiry with devices like Google Home. In 5-10 years, I expect there will be a greater interconnection between our day-to-day applications and services, going beyond calendar management into automated phone calls and appointment scheduling, creation of meal plans, and shopping lists that can be integrated with grocery delivery services, etc.
Ultimately, I expect a greater adoption of voice-activated services that allow us to reduce hands on keyboard or hands on phone time.
IAA AITF WORK STREAM: INNOVATION
Q: What got you interested in AI?
A: Movies! I watched I, Robot probably one too many times when it came out in 2004. Later, other movies such as Sam (voice companion), Transcendence (ingenious scientist creating a sentient machine) and The Entity in the most recent Mission Impossible franchise, all made me contemplate the seeming inevitability of technology singularity.
So, I set out to learn more about AI and machine learning. I completed several specializations offered by Deeplearning.AI via Stanford University. When I was asked to script and film content for the AI track for the Society of Actuaries’ Online Learning Center, I was thrilled to have the opportunity to combine my creative side (plots of all these movies came rushing back to me), with the knowledge I had gained through academic studying. I look forward to the AI track being launched later this year!
Q: Why is getting involved in this initiative/topic important to you?
A: It was intuition. Although I haven’t been an early adopter of technology previously – I never lined up overnight for the latest iPhone release, nor was among the first wave of Blue-Ray or QLED TV owners – I felt an instinctual nudge when I thumbed my first conversation with ChatGPT 4.0, a crystal-clear recognition that I must get on top of the generative AI development. Getting involved with AI has already drastically changed my life, and it will alter the future of work for nearly everyone in society.
Q: How do you expect to interact with AI in 5-10 years?
A: Well, I already interact with AI much more today than I could have imagined just a year ago. Today, I can build and launch a website with AI’s help in two hours, I can banter with Meta AI (via WhatsApp) to stay sharp and snappy, I can character-train AI to act as a coach, and I can use AI as a sparring partner to test out new business ideas.
In 5-10 years, I suspect what we could use AI for is only limited by our imagination. I expect artificial general intelligence (the type of AI that can perform just as well or better than humans on cognitive tasks) will be exponentially closer to reality if not already widespread. I find the plausible future of being able to leverage much more powerful tools to be extremely exhilarating. (Imagine telling someone who just debarked the Mayflower after 66 days of agonizing travel that in a mere 400 years, their journey would be replaced by a seven-hour, flying-in-the-air machine with a lie-flat bed. It would be challenging for that person to wrap their head around it.)
