AI, but not AI

Editor's Note Dhirendra Kumar’s insights and timeless advice for investors --------------------------------------------------------------- 21-September-2024 --------------------------------------------------------------- Dear {NAME}, Every Saturday, I share my perspectives on a topic investors will find useful. This time I delve into what shape should automated, machine-driven financial advice take. AI, but not AI Robo-advisors are an inevitability. Let me qualify that a little bit. I'm talking specifically about general personal finance advice and mutual fund investment advice, which are almost certain to be increasingly delivered by robo-advisors. When I say robo-advisors, I mean an automated, software-driven replacement for a real human advisor. However, there's a caveat to this that will surprise most of my readers: By robo-advisors, I do not mean 'AI', at least not in the sense that the term is used today. Specifically, I do not mean AI driven by large language models (LLMs). Let me explain the two points I have made. A vast majority of mutual fund investors need someone's advice. This is especially true when they are starting to invest. However, at that stage, the money they are investing is simply not enough to make it economical for them to receive personal, human advice from a high-calibre, competent advisor. This creates a significant barrier to entry for many potential investors, particularly those from middle-class backgrounds or young professionals just starting their careers. The economics simply don't add up – high-quality human advisors charge fees that cannot be justified for small investment portfolios. As a result, many novice investors are left to navigate the complex world of mutual funds on their own, often leading to poor investment decisions or, worse, avoiding investing altogether. This is where robo-advisors step in, filling a crucial gap. They can provide standardised yet personalised advice based on algorithms and predefined rules, which can be incredibly useful for novice investors. Moreover, these robo-advisors can operate at scale, serving thousands of clients simultaneously, making them a cost-effective solution for investors and financial institutions. As the Indian market matures and more people seek investment advice, we can expect to see a proliferation of these automated advisory services, democratising access to financial guidance. Which brings us to the so-called 'AI'. Regardless of who or what is giving advice, it has to be transparent, understandable and deterministic. Let's say the advice says, for example, to invest 30 per cent of your money in gold. There must be clear and understandable logic stated in the system's rules. Moreover, for transparency and compliance with regulations, the advisory entity should be able to explicitly state why a particular piece of advice was given. These requirements cannot be reconciled with the way that LLMs operate. Instead, an automated advisory system should emulate a human expert by drawing its inferences and conclusions from a set of rules that are explicitly stated and based on what human experts would do. Those who know the history of AI must be smiling right now because what I have just stated – intentionally – is a description of an 'Expert System', which is what AI was supposed to be earlier. An expert system is a type of AI that mimics the decision-making process of a human specialist in a particular field. Wikipedia tells us that they were developed in the 1970s and gained popularity in the 1980s, among the earliest AI technologies to achieve practical success. During this period, they were widely seen as representing the future direction of AI research and applications. Note that an expert system's expertise is domain-specific, just like a human expert's. The funny thing is that current robo-advisory and similar systems map quite closely to what an expert system is, and the better-designed ones work quite well within that methodology. There are many expert systems in investing, even if most use that term. From spreadsheets to Value Research's new [Fund Advisor]( system, the definition fits many systems, even though none are called AI today because of the term's meaning shift. The future of mass financial advice and guidance in India should be built around these rule-based, transparent robo-advisors rather than opaque AI systems. While LLMs and other cutting-edge AI technologies may have their place, the critical nature of financial decisions demands a more accountable approach. --------------------------------------------------------------- Thank you for being a Value Research Insider. I hope you found this note useful and interesting. What did you think of today’s note? [Let me know](mailto:editor@valueresearch.in). If you know anyone who would enjoy it, please forward this email. They can sign up for free [here](. You can also subscribe to the Hindi version [here](. Was this email forwarded to you? [Sign up here]( [vro-logo]( Copyright © Value Research India Private Limited 2024. All rights reserved. C-103, Sector 65 Noida, 201301. This notification mail has been sent to you at {EMAIL} because you are a member of Value Research Online. [Manage Newsletters]( [Unsubscribe]( [Privacy Policy]( Follow us: [twitter-icon]( [facebook-icon]( [youtube-icon]( [linkedIn-icon]( [instagram-icon](