By Paul Nixon Head of Behavioural Finance at Momentum Investments
Artificial intelligence (AI) research lab, OpenAI has shoved us into the future with ChatGPT. The innovation is deceptively simple – the AI is trained to learn and understand exactly what humans mean when they ask questions and respond in a conversational format. The mind-bending piece is that the first public or user version of this tech is near flawless or at least appears that way. It is super-fast and responds in a manner that is almost unsettling in its humanity. In fact, it is near impossible to tell if the answers provided to the question posed are human or AI-generated.
After scratching the surface, however, a deeper and more revolutionary value becomes clear. The AI is sophisticated enough, in seconds, to evaluate and link complex ideas, concepts and even philosophies providing the essence of these linkages to searchers that will reduce search costs dramatically and add value through progressively speedy learning.
This is significant progress.
This is the essence behind the “Medici Effect,” which if you ask ChatGPT will quickly tell you, “The Medici Effect is a term coined by Frans Johansson in his book of the same name. It refers to the phenomenon of breakthrough innovation and creativity that occurs at the intersection of diverse fields, disciplines, and cultures. The term is inspired by the Medici family, a powerful and influential banking family during the Renaissance in Italy, who supported and brought together artists and thinkers from diverse fields, leading to the flourishing of art, science, and technology during that period.”
As a behavioural scientist for example, I am very interested in how we get people to save more and particularly considering the cultural diversity in South Africa. Asking ChatGPT, “How do we get people to save more by taking anthropology into account?” The AI quickly (seconds) gives a concise five-point plan of the key considerations in this problem including cultural context, social norms, defaults, financial literacy, and even more complex ideas around social identity. Instead of Googling, downloading academic papers and/or reading only slightly relevant blogs I have a concise plan generated in seconds exactly relevant to my question that incorporates psychology, economics, sociology, and behavioral science. I doubt this will be replacing Google any time soon, but it definitely replaces the above process that Google solved and as we know the first rule in behavioral science in changing behaviour is, “Make it easy.” This is very easy.
Because with us, investing is personal.
Our innovative research has one goal in mind –
helping you help your clients reach their personal financial goals. Because with us, investing is personal.
Research is important for innovation, which is essential to growth, in generating ideas for new products and solutions, and in the advancement of existing solutions in the investments business.
The Research Hive was set up within Momentum Investments to provide subject-matter expertise on trends and industry-specific insights. This includes our leading research on behavioural finance,
to help you with those important conversations with your clients. When you see our Research Hive logo it indicates that the research is endorsed by our team applying cutting-edge research.
Ultimately, the aim of our research is to effectively help you help your clients reach their personal financial goals. Because with us, investing is personal
Momentum Investments is part of Momentum Metropolitan Life Limited, an authorised financial services and registered credit provider (FSP 6406).
Now let us take it one level deeper. I know as a behavioral scientist that machine learning (ML) is going to be critical in my field (behavioural finance), but I want to know how to use it to help people save more and I can also clumsily include something about the actual programming code involved. So, I type in, “How can I use machine learning to help people save more with code?.” The first two points it comes back with a cut to the chase, fast. We can help people to save money by predicting their behaviour (before they spend) and we can increase personalisation using machine learning, which speaks directly to the concept of personalised nudging. A second to get to the crux of a complex question because these two things are exactly how machine learning should be used in behavioural science, to change this behaviour.
It then proceeds to provide the key machine learning distinctions I need to consider and… it provides me with the likely programming code I will need to use to execute the ML with a handy “copy code” widget. Now of course the devil here lies in preparing the data, but it is clear how much time this is going to save everyone by cutting through a vast amount of clutter.
My wife is a molecular geneticist working in the pharmaceutical industry and, as the marketing manager, one of her core responsibilities is making sure her brand managers convey the science (clinical trials) to the company sales representatives effectively in campaigns. One of her biggest challenges has been getting people to understand the science fast, usually involving extensive reading and research.
Naturally, this technology has already added tremendous value. It took Facebook 10 months to reach 1 million subscribers and Instagram 2.5 months to get the same number. It took ChatGPT 5 days. Imagine a world where the AI has access to your voice calls for training and where we could be heading?