How actuaries and generative AI can work together

Artificial intelligence (AI) has emerged as a transformative force, revolutionising various industries across the globe. One of the most promising branches within AI is generative AI, which has the potential to reshape the way we, as actuaries, approach our work.

The inception of AI can be traced back to the 1950s, when pioneers like Alan Turing and John McCarthy first envisioned machines that could mimic human intelligence. Over the following decades, AI experienced significant breakthroughs and paradigm shifts. From early expert systems to the advent of machine learning, the capabilities of AI have expanded exponentially.

Generative AI technical explanation

Machine learning encompasses a wide range of algorithms with the aim of making decisions or predictions. Generative AI is a subset of machine learning, which uses existing data and the patterns within it to create new content, rather than purely analysing the data. It involves training models to generate original and meaningful outputs, such as images, music, text and even human-like conversations. This breakthrough has opened doors to remarkable possibilities, propelling AI from being an analytical tool capable of automating simple tasks, to a creative force, especially in the actuarial world.

Generative AI uses a neural network called generative model, which learns the patterns and statistics of a given data set and then uses that knowledge to generate new examples similar to the original data. One popular generative model is the generative adversarial network (GAN), which consists of a generator and a discriminator. The generator takes random input and tries to create realistic data samples, while the discriminator learns to distinguish between real and fake data. These two networks compete against each other, leading the generator to improve its ability to generate data that is difficult to differentiate from real data.

ChatGPT example

There are a huge range of AIs specialising in different areas — one AI which has broken into the mainstream media over the past few months is ChatGPT. ChatGPT uses an advanced language model developed by OpenAI which leverages deep learning techniques to engage in human-like conversations, providing informative responses across a variety of topics. It assists users by generating text-based interactions and offering valuable information. ChatGPT broke records, reaching one million users within five days and becoming the fastest-growing app in a short span of time, which shows the ever-increasing popularity of AI.

We have experimented with ChatGPT and the possibilities which it could provide became quickly apparent. Suppose you wanted to create a calculator, which priced a basic insurance product but had little to no prior coding experience. Entering simple inputs into ChatGPT such as “Write a Python code which [insert basic description of required model]” will output lines of code which can be copied and pasted. Simple tweaks and adjustments can be made to the code by asking follow-up questions in a conversational manner. If any errors occur, we found that ChatGPT will often be able to recognise these and adapt its previous code. This is a basic example of what is possible with just one AI, but the prospects of what it can achieve and the time it could save are exciting.

Actuarial use cases

For us actuaries, who specialise in analysing risks and making data-driven decisions, generative AI offers a wealth of opportunities. The ability to simulate vast amounts of data and generate synthetic scenarios will allow us to explore complex risk scenarios more efficiently than was previously possible, develop robust models, and enhance decision-making processes. It enables us to gain deeper insights, anticipate emerging trends, and devise innovative strategies to mitigate risks.

Currently we have not seen clients actively using generative AI in their projects, however there is a hot market for it. There is a lack of AI expertise within the actuarial industry and clients are seeking out consultancies to figure out how they can utilise AI for solution development and to accelerate their tools. A major limiting factor right now is the number of hours in a day – with so many potential use cases and AI currently available, there are many avenues to explore for any given project. Furthermore, care needs to be taken when using generative AI tools as often the data entered into them by users is not private, which could lead to data safeguarding issues for clients – certain AI companies have private keys available which would allow tools involving sensitive data to be modelled. Once a portfolio of successful projects has been developed for the insurance industry using AI, this can be used as a framework for future endeavours.

A hypothetical scenario in which AI could become useful in the actuarial industry is given below:

Risk modelling and stress testing

We play a crucial role in assessing and managing risks for insurance companies, pension funds, and other financial institutions. AI can revolutionise risk modelling by simulating thousands of potential scenarios and stress testing portfolios to evaluate their resilience. By training AI models on historical data, we can generate data that accurately represents real-world scenarios. This will enable us to assess the likelihood and potential impact of catastrophic events, such as floods, hurricanes, or market crashes, on their clients' portfolios.

AI models could be used to simulate the potential impact of climate change on insurers' exposure to flood risks. By generating synthetic flood event data based on historical patterns and projected climate scenarios, we can estimate potential losses and assess the adequacy of reserves which will in turn allow us to inform pricing decisions more accurately.

Conclusion

Generative AI has the potential to revolutionise our work as actuaries and reshape the insurance industry. We have only begun to scrape the surface of what is capable with the current AIs at hand, and the technology is only going to improve at a rapid rate. It is an exciting space to watch and the best time to get involved in this modern technological movement is now.

The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organization or its member firms.