Ask ChatGPT where to invest $10,000, and it’ll give you a confident, articulate answer — not a trustworthy one. Whilst it may sound like a portfolio manager, it is key to understand that it is not one.
Still, that illusion is spreading fast. A survey from trading and investing platform eToro of investors in the US found that 58 per cent are adopting AI tools, such as ChatGPT, to build their portfolios.
Large language models (LLMs) have changed how we access information, yet they are linguists, not financial experts. In a new era of investment management revolutionised by AI, they will not be the ones picking the stocks, but they may still be the face of this transformation.
Don’t mistake fluency for financial literacy
ChatGPT will not be your investment manager and for good reason. LLMs are trained to predict language, not financial outcomes. They can’t process live market data, calculate risk-adjusted returns, or optimise portfolios with mathematical precision. In a sector where accuracy and accountability are vital, that’s a fatal flaw.
Even if the technology could make sound trades, regulators still have the final say.
The US Securities and Exchange Commission and Finra demand explanations, with the latter issuing a regulatory notice in June 2024, stating that the same securities rules apply when using LLMs. A system that cannot fully justify a trade would immediately raise alarm bells for regulatory bodies.
Investment advisers are already falling foul of this. In March 2024, the SEC brought its first ‘AI-washing’ enforcement, which saw two advisers charged for making misleading statements about their use of AI in investment decision-making.
The future of AI in investment
Scrutiny around LLMs does not mean AI has no place in investment management. It already does, just not the type that helps kids with their homework.
Quantitative AI systems already outperform humans in certain areas. The key difference is that these models are transparent. They are back-tested against decades of market behaviour and designed to be auditable. These systems meet regulatory and fiduciary standards and are steadily being embedded into robo-advisers and digital wealth platforms.
This is not to say that LLMs are completely useless. The possibility of incorporating LLMs into a full AI management solution does exist. LLMs’ strong suit is their linguistic skills, so they can translate user inputs for the quantitative AI system and vice versa.
Imagine telling your AI adviser, ‘Keep my exposure to tech under 50 per cent and rebalance quarterly’. The LLM understands the request, translates it into machine-readable commands, and then passes it on to the specialised quantitative model to execute.
The same model then feeds results back to the LLM, which explains your new portfolio in clear, conversational terms. In this sense, LLMs become the translators, bridging the knowledge gap between the quantitative model and the users themselves.
The generational shift
To older investors, an AI adviser may feel cold or impersonal. To Gen Z, who will be the dominant generation of investors in 10, 20, or 30 years’ time, it feels natural.
A Walton-GSV-Gallup survey found nearly 80 per cent of Gen Z in the US have used AI tools, mostly LLMs, with this number only set to rise. These tools aren’t only being used for entertainment but also for help with decision-making.
As this generation accumulates wealth, they will expect financial advice to be conversational, personalised and always on. To them, a chat-based adviser won’t feel futuristic, it will just feel normal.
It is safe to say LLMs will not be managing portfolios, they are just not built that way. However, they will be the faces of the systems that are. The future of investment management will have multiple layers. Language systems that interpret and explain, with quantitative engines that decide and execute.
To the investor, though, it’ll all look like one thing: a conversation with their new AI adviser.
These systems currently operate independently, so a set-up like this has not quite arrived. For now, at least, AI is not yet coming for investment advisers’ jobs. However, it does provide a key opportunity to get ahead of the pack.
Many advisers still rely on fragmented legacy tools that quietly erode efficiency and control. New AI solutions are giving advisers a holistic view of their portfolios, enabling them to automate account management and personalise solutions for each individual client while saving time and money.
There is still a large premium for human interaction, so those advisers who learn to leverage this new technology as quickly as possible, while maintaining that person-to-person contact, will come to rule the roost.
Christopher Ainsworth is chief executive of Pave Finance
