Crypto, Cholesterol and ChatGPT – Why AI is nowhere near replacing your wealth advisor
Dominique Jooris
C-Suite Private Wealth and Family Office | Wealthtech | Global Investment Banking | Author & Public Speaker
“You are making our job obsolete” exclaimed a former colleague turned wealth manager when she sat down for a demonstration of the latest release of the AI-driven insights of our wealth advisory platform WMCockpit.
As much as I was flattered, the truth is that while AI can deal with an ever-growing share of human behavioral systems, it remains limited in its ability to deal with our individuality.
The relevance of AI results to a specific client is hampered by two factors or biases: the source material available to scour and draw data from on one hand, and the ultimate complexity of individual decision on the other.
We have all been brainwashed in believing that high cholesterol leads to cardio-vascular diseases.??AI will provide you with the best estimate of what “too high” is, drawn from information scrubbed from reports and studies heavily influenced by the pharma industry, and only a sprinkling of dissenters – the first bias.??
Should you go with the flow, medicate your condition and then experience side effects, AI will also indicate which drugs can mitigate these.??Now, new studies have linked cholesterol-lowering statin drugs with a higher incidence of Alzheimer.??
Your choice of using drugs to correct hypercholesterolemia, despite leading to statistically quantifiable outcomes, will be driven by your experience, emotions and humanity, rather than numbers.??Witnessing a distant relative dying on the operating table during a bypass operation (incentive to medicate), or instead having a parent drifting into dementia (incentive to avoid drugs) will influence your decision and trump your physician’s analytics showing that the risks of the former are much larger than those of latter – the second bias.
Enough of health management, back to wealth management.
Who are the people writing papers on Crypto???Traders, research analysts, journalists covering that eco-system, individuals linked to exchanges and lobbyists paid to further the agenda of exchanges (FTX anyone?)??- In one word, a lot of stakeholders.??How many people wrote papers to support Jamie Dimon’s view expressed this year[1], that Bitcoin is a “hyped-up fraud, it’s a pet rock”???Neither him nor Warren Buffet will clog-up the internet with dissertation about what is evident to them.??See the bias right there?
领英推荐
When WMCockpit fed the question of optimal crypto allocation to ChatGPT, the number ranged from 5% for estates of USD5 mm to 10-20% for estates of USD100 mm and up to 40% for estates above USD500 mm.??Any private banker suggesting the latter numbers to their clients is buying a one way ticket to an psychiatric asylum where they would be expected to attempt enrolling stony-faced nurses into Valium-trading Ponzi schemes.??
Why such high numbers though???Because the question implicitly asked from a mathematical perspective (historical and expected returns, correlations, diversification) does not factor in the perceived risk of fraud, wallet hacks, regulatory clampdowns, or lack of crypto-denominated productive assets to name a few.
As individuals, we intuitively “know”[2]?that these numbers are too high, but the fact that they are produced by AI cannot be ignored.??Someone who has witnessed a close acquaintance losing her life savings to a bad crypto trade will be nudged on the cautious side of the decision.??Had the same person witnessed a crypto tale of rags to riches,??you will likely see them chomping at the bit, FOMO and all.
Whatever the logic or systemically optimistic perspective behind the AI result,??values like this should challenge one’s complete skepticism about cryptocurrencies, because they speak of widespread acceptance, irrespective of whether the actual usage is pervasive.
In the field of wealth management, I believe that?Analytics?will continue to provide a black-and white Cartesian mathematical framework.?AI?will drive metrics that can be used as pointers.??Rather than considering these as absolute, they will remain the expression of a fuzzy logic[3]?process, but capture a much broader dataset than what is accessible through Analytics alone.??Each individual client will and should then apply his or her?Human Overlay?to the decision process and get to the right allocation.??
These three perspectives are and will remain complementary and ensure a bright future for wealth managers.??I would raise my glass to that, but my cholesterol is too high.
[1]?Jamie Dimon, on CNBC’s Squawk Box, Jan 19, 2023
[2]?A relevant reading here: “Blink” by Malcolm Gladwell
#wealth
I help senior leaders in transition increase their impact, influence, and happiness | x-Goldman Sachs | Author | I speak 5 languages and eat kimchi with everything (including oatmeal!) | Canada, Hong Kong
10 个月Thank you for the insightful commentary. People tend to forget the 2 (among many) important points you highlight here: 1. AI is as biased as the data it bases its output on. 2. Much market information is written by the very people who are trying to make money of said financial product and therefore is inherently biased. Hooray for wealth advisors!