Guidance Resource: Artificial Intelligence and Discrimination in Insurance Pricing and Underwriting

2 CPD Points

Data Meet Up

The Sydney AI and Data Science Meetup featured a presentation by Chris, an actuarial recruiter at SKL Technology, who shared insights on the job market and career opportunities in AI and data science, adapting his previous year's presentation to reflect recent changes. The discussion covered the evolving landscape of data science and AI roles, including job search strategies and the importance of self-reflection to align skills with suitable positions, while highlighting the diversity of tasks and backgrounds within the field. The conversation ended with insights on recruitment practices, employer strategies, and career opportunities for junior roles and fresh graduates, emphasizing the value of practical experience, networking, and ethical considerations in AI.

Chris Gray

2 September 2025

Submission

1 CPD Point

Submission to Treasury: Economic Reform Roundtable

The Institute has provided policy reform proposals to Treasury's Economic Roundtable consultation in the areas of climate adaptation and resilience investment, AI governance, quantum computing preparedness, tax reform, superannuation and financial advice and guidance, the health system and improved data management.

Actuaries Institute

24 July 2025

Video

2 CPD Points

Data Meet Up

This meeting focused on the role of mathematics in AI, particularly in building large language models, with discussions on tokenization, embeddings, attention mechanisms, and training methods. The speakers explored the capabilities and limitations of AI models, including their ability to handle various tasks and languages, as well as the challenges in implementing AI across different sectors. The conversation ended with considerations on the growing demand for AI and its potential overuse, emphasizing the importance of balancing AI implementation with customer needs and maintaining human oversight in areas like coding and data science.

Sachin Abeywardana

15 July 2025

Video

2 CPD Points

Data Meet Up

The meeting covered various aspects of AI and data science in education and professional settings, including discussions on teaching methods, challenges with AI-generated content, and the use of AI tools in course development and feedback processes. Participants explored the integration of AI in workflows, the importance of human oversight in AI systems, and the concept of instinctive recognition of flawed models. The event concluded with announcements about job opportunities and consultancy services, as well as an invitation for networking.

Matthew Beauregard

3 July 2025

Media Release

0 CPD Points

Andrea Cross announced 2025 Women Leaders in AI and Data Science Scholarship winner

The scholarship, a joint initiative between the Actuaries Institute and Chief Executive Women (CEW), recognises Ms Cross's exceptional leadership across more than two decades in technology roles spanning innovative start-ups and senior government positions. As CIO of the Australian Government’s lead economic advisory body, Ms Cross brings a unique perspective to the intersection of artificial intelligence, data science and public policy.

Actuaries Institute

30 June 2025

Video

2 CPD Points

Data Meet Up

The meeting covered various aspects of AI and data science in education and professional settings, including discussions on teaching methods, challenges with AI-generated content, and the use of AI tools in course development and feedback processes. Participants explored the integration of AI in workflows, the importance of human oversight in AI systems, and the concept of instinctive recognition of flawed models. The event concluded with announcements about job opportunities and consultancy services, as well as an invitation for networking.

Luana Woodward

24 June 2025

Papers

2 CPD Points

Summit 2025: Mulplicave SHAP Values: Advancing Interpretable Machine Learning in General Insurance Pricing

Insurance pricing increasingly relies on machine learning models who contradict actuarial pricing’s multiplicative structure. This paper extends SHAP values to multiplicative decompositions, aligning analytics with traditional actuarial rate relativities.

Jamie Wu

24 June 2025

Presentation slides

1 CPD Point

Summit 2025: Practical Learnings in applying GBM, GLM and traditional analysis in modelling Life Insurance Claims and Lapses

Recent advancements in data analytics within the life insurance sector have seen a growing adoption of linear modelling techniques, for example Generalised Linear Models (GLM) and LASSO Regressions. This presentation aims to extend the exploration by evaluating the performance of non-linear techniques, such as Gradient Boosting Machines (GBM), in comparison to linear and traditional methods. We will also address practical considerations for actuaries when applying these methods in typical life insurance analyses. Our presentation will focus on the application of GBM, GLM, and traditional analysis methods in the context of claims and lapse analysis to: • Assess the relative performance of these techniques, including variations in performance based on the type of analysis conducted. • Identify properties of these techniques that are beneficial upstream and/or downstream, including methods for interpreting model outputs • Investigate potential methods to explain differences in predictions among the tested methods. To achieve these objectives, we will utilise the following datasets: • Claims – New Zealand Industry Disability Income Experience Investigation Data • Lapses - Synthetic dataset created by IBM on customer churns for a Telecommunication company

Timothy Lam and Jin Cui

13 June 2025

Presentation slides

1 CPD Point

Summit 2025: AI Retirement Information Centre / HelpDesk

Financial advice for retirement planning is currently burdened by excessive red tape and high costs, making it inaccessible for many average superannuation members. As a result, financial advice is often sought only by high-income earners, leaving those who may need it most without support. While there are many guides available online, they can be difficult to navigate, are not tailored to the specific questions members have, and are often outdated due to frequent changes in superannuation and tax rules. In some cases, these guides provide contradictory information, further reducing trust and quality. This presentation aims to explore ideas and facilitate discussions on using AI to provide scenario-based retirement planning information ("not advice") to members. Some ideas on how this may work include: Using AI to analyse member-provided information and categorise members into different cohorts. Based on cohort characteristics, using AI to identify common retirement-related questions and generate relevant scenarios. Ensuring AI inputs are always updated with the latest rules from official sources for accurate calculations. Generating retirement outcome results based on selected scenarios, along with a list of key assumptions used. Allowing members to adjust key assumptions for "what-if" scenarios, providing revised retirement outcomes. The goal is not to replace professional financial advice but to offer up-to-date, scenario-based retirement information tailored to members' specific questions and needs.

Edward Tam and Abrren Chen

13 June 2025

Podcast/Vodcast/Audio

2 CPD Points

Summit 2025: Solving for Underinsurance – Serving the Underserved Younger Generation

The younger generation remains underserved in the life insurance space, presenting significant opportunities for market growth. Effective strategies must address regulatory and structural barriers, while focusing on distribution, product fit, and perceived value. Recent regulatory and environmental shifts have created a more supportive landscape for life insurers to engage this segment. Targeted distribution models, aligned with key life moments, can help insurers shape perceived value and encourage early engagement — building trust and establishing long-term customer relationships. A seamless digital experience is essential for both acquisition and retention, offering opportunities to innovate across the entire value chain. Insurers must continually analyse and adapt to evolving customer expectations, leveraging data to personalise offerings and strengthen engagement. Success will require insurers to be agile, data-driven, and proactive in navigating emerging challenges. The speakers will provide a view as to how technology and data-driven thinking can provide impetus and opportunities for life insurers to shift and evolve how they address the emerging insurance needs of the younger demographic and remain affordable, relevant, and sustainable for the community.

Alan Merten and Marc Mer

13 June 2025

Podcast/Vodcast/Audio

2 CPD Points

Summit 2025: Practical Learnings in applying GBM, GLM and traditional analysis in modelling Life Insurance Claims and Lapses

Recent advancements in data analytics within the life insurance sector have seen a growing adoption of linear modelling techniques, for example Generalised Linear Models (GLM) and LASSO Regressions. This presentation aims to extend the exploration by evaluating the performance of non-linear techniques, such as Gradient Boosting Machines (GBM), in comparison to linear and traditional methods. We will also address practical considerations for actuaries when applying these methods in typical life insurance analyses. Our presentation will focus on the application of GBM, GLM, and traditional analysis methods in the context of claims and lapse analysis to: • Assess the relative performance of these techniques, including variations in performance based on the type of analysis conducted. • Identify properties of these techniques that are beneficial upstream and/or downstream, including methods for interpreting model outputs • Investigate potential methods to explain differences in predictions among the tested methods. To achieve these objectives, we will utilise the following datasets: • Claims – New Zealand Industry Disability Income Experience Investigation Data • Lapses - Synthetic dataset created by IBM on customer churns for a Telecommunication company

Timothy Lam and Jin Cui

13 June 2025

Podcast/Vodcast/Audio

2 CPD Points

Summit 2025: AI Retirement Information Centre / HelpDesk

Financial advice for retirement planning is currently burdened by excessive red tape and high costs, making it inaccessible for many average superannuation members. As a result, financial advice is often sought only by high-income earners, leaving those who may need it most without support. While there are many guides available online, they can be difficult to navigate, are not tailored to the specific questions members have, and are often outdated due to frequent changes in superannuation and tax rules. In some cases, these guides provide contradictory information, further reducing trust and quality. This presentation aims to explore ideas and facilitate discussions on using AI to provide scenario-based retirement planning information ("not advice") to members. Some ideas on how this may work include: Using AI to analyse member-provided information and categorise members into different cohorts. Based on cohort characteristics, using AI to identify common retirement-related questions and generate relevant scenarios. Ensuring AI inputs are always updated with the latest rules from official sources for accurate calculations. Generating retirement outcome results based on selected scenarios, along with a list of key assumptions used. Allowing members to adjust key assumptions for "what-if" scenarios, providing revised retirement outcomes. The goal is not to replace professional financial advice but to offer up-to-date, scenario-based retirement information tailored to members' specific questions and needs.We will unpack the design of MyNorth lifetime through several case studies, and feedbacks from advisors and members who took up the solution. We will also explain the key features that boost retirees’ confidence to drawdown at higher rates. Our average MyNorth Lifetime client has improved their income by ~50% through a combination of high Lifetime rates, higher age-pension eligibility, and increased drawdown from account-based pensions. The majority of Lifetime members no longer draw the minimum from their account-based pensions when this was previously not the case. Over FY24, we saw lifetime members drawing down on average 63% higher income than the statutory minimum rates for age 60 to 64 group, and the uplifts were 54% and 41% higher for age 65 to 74 group and age 75 to 79 group respectively. This reflects members expressing a willingness to spend more freely and enjoy greater retirement living standards once they have overcome the fear of running out of savings. In addition, we will also provide a high level overview of the implementation roadmap of our Retirement Income Strategy including other building blocks to achieve the objectives and further considerations to help boost retirees’ confidence to enjoy the best possible retirement they deserve to have.

Edward Tam and Abrren Chen

13 June 2025

Presentation slides

1 CPD Point

Summit 2025: Increasing the Per Capita Output of Actuaries to grow the Profession to 2030

This discussion aims to estimate the per capita output of the actuarial profession over time and to argue that, in order to expand its reach, this output must increase significantly. How can this be achieved? We need to approach this objective in two ways simultaneously. First, actuaries in traditional industry sectors such as Life, General Insurance, Superannuation, or Investments must deepen their expertise and thus offer a wider range of services within these areas. Second, we should explore new fields and apply actuarial skills to address emerging challenges. The critical question is how the relatively small actuarial profession can enhance its per capita output to extend its reach within existing business lines and create the capacity to enter new fields. With recent technological advancements, such as robotics and LLMs into society, how might we leverage AI to help fulfil the aspirations of the actuarial profession. However, what implications does this have for job security? Will we be replaced by robots and Super Intelligence? As AI becomes more prevalent, what will the new ecosystem look like? How will the actuarial operating model evolve? How should we embrace new technology to combat the future uncertainties? We will explore these topics and the evolving role of actuaries in a TED talk scheduled for June 2025.

Richard Yee, Joseph Hoang-Luu and Steven Choi

12 June 2025

Presentation slides

1 CPD Point

Summit 2025: Modernising Actuarial Practice to Address Climate Change and Societal Issues: A Systematic Literature Review

Technology, data science, artificial intelligence (AI), climate change, and other societal issues drive a revolutionary period that the actuarial profession must navigate. Actuaries are unusual in that they can tell stories with data, allowing for better understanding and major benefits for business, government, and society at large. Actuaries must adapt to changing times and use new technologies to remain relevant and uphold their trusted advisory function, given the complexity of today's dangers and possibilities. To address urgent problems like climate change and global uncertainty, this research examines how actuaries might use new methods to adapt to this disrupted environment. This study aims to evaluate the approaches that can fulfill the needs of modern world and determine the best ways to solve societal issues like climate change and demographic transitions while integrating technical advancements like artificial intelligence (AI) and big data into actuarial practice. The study also aims to investigate how actuaries can continue to steer well-informed decision-making processes that influence the future and balance the interests of various stakeholders. A thorough analysis of the research, case studies from various industries, and a systematic literature reviews using PRISMA used for research approach. These techniques evaluate the effects of developing technologies on risk management, financial modeling, and decision-making processes and provide a comprehensive understanding of their practical applications. The results imply that actuaries can better offer creative answers to contemporary problems when they collaborate with AI, data science, and other emerging technologies. Actuaries also need to continue becoming forward-thinking and adaptable, concentrating on social demands like sustainable development and climate resilience. Actuaries can continue to provide value in a world that is changing quickly if they can turn uncertainty into opportunity, and emphasize the profession's role in creating a resilient future.

Ingga Martatino, Adrian Mak and Wiesye Santoso

12 June 2025

Presentation slides

1 CPD Point

Summit 2025: AI, Education, and Actuaries

Artificial intelligence (AI) is poised to revolutionize the actuarial profession, making it imperative to establish best practices for its integration into actuarial education. To ensure that future actuaries are well-prepared for the evolving demands of the field, it is essential to incorporate AI both as a powerful tool to enhance learning and as a critical component of the curriculum. Best practices should focus on training actuaries to use AI efficiently and ethically, harnessing AI to improve education delivery and assessment, and addressing AI's ethical and professional implications in actuarial work. By proactively integrating AI, actuarial education can be aligned with the technological advancements shaping the profession. Educating actuaries to use AI effectively requires a strong emphasis on core actuarial skills such as ethics, communication, and risk management. As AI systems increasingly influence decision-making processes, actuaries must be equipped to understand and manage the ethical considerations and potential biases inherent in these technologies. Additionally, effective communication with stakeholders about AI-driven decisions is crucial. There is also a greater need to foster critical thinking skills, enabling actuaries to evaluate AI outputs critically and ensuring that these outputs align with sound risk management principles and ethical standards. This approach will help actuaries maintain their role as trusted professionals who can navigate the complexities introduced by AI. AI can significantly improve outcomes in broader actuarial education by serving as a versatile tool across various learning and assessment areas. AI can be utilized as a writing and communication aid, a coding assistant, a feedback mechanism, and a detector of cheating and plagiarism. Moreover, AI can simulate stakeholder feedback through role-playing scenarios, enhancing students' real-life communication and decision-making skills. AI-driven grading tools can streamline assessments, while AI education chatbots offer opportunities for students to practice recall, creativity, and Q&A communication skills, providing real-time feedback on their practice answers. To fully leverage these benefits, it is essential to redesign assessments to focus on human effort, creativity, and critical analysis rather than mechanistic tasks and rote memory, ensuring that the educational process remains relevant and effective in an AI-enhanced landscape. This paper and presentation propose best practices to address the changing educational needs of Australian actuaries. The presentation will include a hands-on demonstration of AI in actuarial education.

Colin Priest

12 June 2025

Presentation slides

1 CPD Point

Summit 2025: Towards Fairer Retirement Outcomes: Socio-Economic Mortality Differentials in Australia

The Australian superannuation system faces the challenge of delivering equitable and sustainable income in retirement, yet the development of effective longevity products remains limited. A key barrier is the lack of detailed understanding of mortality differentials across socio-economic groups, which complicates the pricing and design of retirement income products. While it is known that mortality varies by factors such as income and education, comprehensive evidence for Australia has been scarce, particularly for the post-retirement population. This paper addresses this gap by analysing individual-level linked data from the Australian Bureau of Statistics’ Personal Level Integrated Data Asset (PLIDA), covering the entire Australian population aged 60–100 over 2016–2017. We examine mortality differentials across socio-economic indicators including area-level socio-economic advantage and disadvantage, income, marital status, and home ownership using flexible Hermite-spline Poisson regression models. Our results reveal substantial disparities in mortality and life expectancy across socio-economic groups, with differences narrowing at older ages. For example, at age 60, the gap in period life expectancy between the most and least advantaged males is 11.5 years, and 9.1 years for females. Longevity differences translate into substantial variation in annuity income. The large and persistent differences by socio-economic characteristics suggest that uniform approaches to longevity product pricing may unintentionally disadvantage certain groups. These findings highlight the importance of accounting for heterogeneity in longevity when designing retirement income products and inform policies aimed at fairer outcomes in the retirement phase.

Fei Huang

12 June 2025

Guidance Resource: Artificial Intelligence and Discrimination in Insurance Pricing and Underwriting

Related publications

Be informed. Stay ahead. Subscribe.

Resources

Qualification programs

The Institute

Follow us