Kah Hong T

Kah Hong T

Quantitative Research and Machine Learning

AI Financial Analyst?

3 minute read

Published:

I have recently read several articles echoing a similar sentiment: due to the limitations of artificial intelligence, AI will not replace traditional financial analysts anytime soon. One author suggests that human intelligence is still decades ahead of AI in automated fundamental investing. However, despite the confidence seemingly portrayed in these articles, the rationales and empirical evidence provided to support this notion have been weak.

There are two main reasons why this conclusion is surprising. First, this pessimistic outlook contradicts the traditionally optimistic, narrative-driven view of equity analysts. Second, a closer look at the empirical research used to support this notion reveals that the conclusions were drawn by surveying traditional professionals in the financial industry who may know little about AI.

Consider these examples:

  1. Pessimistic Outlook by Morgan Stanley
    Morgan Stanley surveyed their investors and financial advisors, finding that customers still “prefer a human touch for the time being,” with “80% of survey respondents saying AI would never completely replace human guidance.”

  2. Why AI won’t replace fund managers and equity analysts anytime soon
    The author presents examples where OpenAI Chat failed to provide research output and concludes that some calculations and insights require “deep industry knowledge” and “significant guidance to produce the desired results.” It is astonishing that the conclusion—that AI replacing human financial analysts is “decades away”—was drawn from this superficial usage of the ChatGPT interface.

The value of AI models lies in the orchestration of various contextual, memory, and reasoning components; a zero/single-shot prompting of models was never going to sufficiently consider contextual information and incorporate enough reasoning capability to generate credible market insights.

I strongly believe it is premature to conclude that AI will never be able to replace humans in making effective investment decisions, especially without a clear understanding of where this technology will go. Furthermore, there is a common argument against the use of AI in financial analysis, which is that the technology is not ready for fully independent investment decision making. But who can we expect the making the technology ‘ready’? Certainly not OpenAI, Anthropic, or Meta, as the foundation models builders cannot be expected to allocate their intellectual or compute resources towards a narrow field such as investment management, or any other other field that is not in their core business. As proclaimed by Elon Musk, technology does not advance by itself, and can only do so through sincere and thoughtful research over long periods of time. It is up to individuals to make AI ready for investment decision making.

Moreover, research institutions or tech-enabled fund managers have little incentive to broadcast the success of any fully autonomous AI agent in trading, as it risks increasing competition and the alpha being arbitraged away. Thus, even if a model were to work effectively, it would not be publicly acknowledged, as it would quickly cease to work in practice.

Current AI research and data platform companies aim to enhance the abilities of traditional, fundamental fund managers. However, if there is a technical case for an orchestrated AI workflow directly leading to investment decisions, the next frontier for AI in finance becomes clear: a fully autonomous AI-driven research analyst, or even a research team.

Now, this could really be an interesting endeavour for machine learning researchers to work on.

*Disclaimer: Opinions expressed are strictly my own, and do not represent any firm or institution.

You May Also Enjoy

Focus on Building

less than 1 minute read

Published:

Took a breather from the myriad changes happening the last few weeks. This year has been eventful indeed. Some principles to follow as I pursue a more definitely uncertain, indefinitely successful and infinitely interesting future.

OpenAI Reasoning Model o1

less than 1 minute read

Published:

OpenAI just released the o1 model that is focused on reasoning capabilities, generally substituting immediate next token generation with some reasoning steps introduced before some delayed token generation. This represents a significant 0-1 step-up in LLM capabilities, immediately surpassing beyond Gemini, Claude and other open-sourced models like Llama3 in mathematical and scientific reasoning. Most likely, this model does not represent a foundational model improvement, but rather entails introducing an orchestration layer informed by chain-of-thought reasoning, self-reflection and reinforcement learning. As opposed to zero-shot prompting using the base pretrained model, this means perhaps a factor more number of inference calls for each request. This has interesting implications on the output token cost structures of using AI models with reasoning.

Competition as Validation

less than 1 minute read

Published:

Peter Thiel: Competition as Validation. It may be tempting to squeeze through the narrow door that everyone is trying to get into, but take a step back and instead look around the corner and explore the vast, empty gate that no one is paying attention to. It is unthinkable that in many perceivably esteemed career paths the degree of competition has become so fierce when the stakes are so low. The intention is not to insinuate a lifestyle lacking in competition, and thus lacking in ambition. While thriving to succeed in a highly competitive space may indicate some level of ambition, the reverse is not true.

Action and Conviction

4 minute read

Published:

Learnt an inspirational quote from Sam Altman today: Inaction is an insiduous form of risk. Only by moving forward can we see and explore the wide universe of opportunities ahead. Have found myself starting to appreciate the complexity of risk management in personal or career decision making; a state of action or inaction both involves accepting a level of risk, and have therefore decided to restructure the problem into 2 concepts: Action and Conviction.