How will AI revolutionise investing?

Since OpenAI released the demo version of its artificial intelligence-powered ChatGPT chatbot in November 2022, the world has been in a frenzy about the opportunities for improvement that generative AI can bring to almost all walks of life. Public usage of ChatGPT shot up and soon after, other similar applications were released by rival companies.

The impact of generative AI will be felt across all business sectors, but some are likely to feel it sooner than others. Investing is likely to be among the first to be transformed by the new technology. The ability of AI to analyse large datasets at speed could bring in new, faster ways of making investment decisions.

Here are some of the main ways in which AI can revolutionise investing:

1. Advanced data analysis: AI algorithms can process massive amounts of financial data and identify patterns that humans may not be able to detect. This can help investors make more informed decisions and spot potential investment opportunities.
2. Automated trading: AI-powered trading systems can execute trades at high speeds based on predefined algorithms. This eliminates the need for manual intervention and can capitalize on market movements in real-time.
3. Risk assessment and management: AI can assess market risks by continuously monitoring various factors, such as economic indicators, news sentiment, and historical data. This enables investors to make more accurate risk assessments and adjust their portfolios accordingly.
4. Personalized investment recommendations: By analysing an individual's financial goals, risk tolerance, and market conditions, AI can provide personalized investment recommendations. This can help investors optimize their portfolios and achieve their financial objectives.
5. Sentiment analysis: AI can analyse social media, news articles, and other sources of information to gauge market sentiment. This can provide valuable insights into market trends, public perception, and potential impact on investments.

Let’s take them in turn to see how they could affect investing as we know it. When it comes to advanced data analysis, AI algorithms have the power to process huge amounts of financial data quickly, and find patterns that humans may not be able to detect. This can help investors to identify potential opportunities that they may otherwise miss.

Identifying new correlations

For example, the AI algorithm could identify a direct correlation between the publication of an obscure economic indicator and stock price movements that investors may have overlooked as the particular indicator is not considered to be important by human market participants. This is already taking place to some extent.

However, investors must be careful and not trust the AI blindly. Historical patterns may not be relied upon to accurately predict future market behaviour, but the algorithm will not know this. Instead, it may confidently predict positive future returns regardless of other factors that influence investors’ decisions, such as geopolitical events or contagion from movements in the prices of other, apparently uncorrelated, assets.

Automated trading is a logical follow-up. Already, high-frequency trading (HFT) systems execute trades at high speeds, based on algorithms that seek to take advantage of any instance of market inefficiency. However, to best take advantage of automated trading, investors should trade large amounts, because HFT systems generally rely on scale in order to deliver meaningful returns.

In the area of risk assessment and management, AI makes a significant contribution to investors’ ability to react quickly to changes in market conditions. For example, it can alert investors to economic indicators overshooting or undershooting expectations, by monitoring data releases, market conditions and breaking news in real time – something that a human cannot do as fast as an algorithm.

AI could also use natural language processing (NLP) algorithms to analyse news articles, social media posts, and other sources of information to identify trends and sentiment related to specific companies or industries. This would help investors make more informed choices than if they did not have this data.

Robo-advisors on the rise

To make personalised investment recommendations, the AI requires inputs such as the investor's risk tolerance, investment goals, and financial situation. The AI considers macroeconomic and microeconomic factors to analyse historical data and predict future performance, and uses machine learning algorithms to make recommendations based on the investor's unique circumstances.

There are already AI-powered tools out there that are making personalised investment recommendations, all with varying degrees of success, which have been dubbed “robo-advisors”. In the UK, perhaps the best known of these is Nutmeg, the online investment platform that offers fully managed portfolios, fixed allocation portfolios or socially responsible portfolios.

Nutmeg disputes the term “robo-advisor,”¹ arguing that it actually offers a blend of technology and human decision-making. Its flagship fully managed portfolios are built using human decisions as to what ETFs make up a user’s portfolio and in what proportions, and it is then managed with the help of AI-assisted information about market movements, news, and data analysis.

Another example is Forbes-backed Q.ai, an online investment platform and app that provides automated financial advice based on powerful deep-learning technology and makes trades on users’ behalf. Users set their risk tolerance and investment horizon and choose between different investment themes (such as growth, value, income), and the AI creates a portfolio matching the user’s preferences and goals. The AI monitors market conditions and makes adjustments to the portfolio seeking to maximise returns and minimise losses.

Finally, there are various ways in which AI conveys its sentiment analysis, but the principle remains the same: using natural language processing (NLP), AI can analyse news articles, social media posts, and other pieces of text to gauge market sentiment. It then communicates its findings in a particular way, for example it presents a binary (positive/negative) result, or comes up with scores on a scale for a more nuanced approach.

Democratising access to opportunities

Used for years by specialist investment strategies such as M&G’s machine learning-backed strategy, AI’s advance into mainstream investing is likely to democratise access to fast, comprehensive analysis of huge amounts of data. This can offer investors opportunities they may have otherwise missed, as well as an earlier look at emerging trends in the markets, and also better diversification.

However, this new dawn of investing is not without pitfalls. Because AI relies so much on past data to arrive at its recommendations, there is the danger that the models will make incorrect predictions, which will lead investors to take the wrong investment decisions. As it is well known, past performance is not a reliable guide to future performance.

Some AI models do not provide clear reasons for their decisions. This lack of transparency can make it hard for investors to understand the rationale behind certain investment recommendations. Investors could therefore be missing crucial information that they need in order to decide whether to invest or not.

There is also the important issue of “hallucination” – when the AI model returns answers that seem correct, but are entirely made up. For example, earlier this year a lawyer in New York submitted a legal brief citing six non-existent judicial decisions produced by ChatGPT. He later apologised for relying so much on the new technology that he did not check its answers were true.

Because of this, it is likely that for now, AI will be used in investing mainly to streamline existing tasks and make them more efficient, rather than being entrusted completely with most of the process. A recent report by Deloitte² lists some ways in which AI could be used to that end:

• Reading earnings transcripts to assess business managers’ sentiment
• Identifying relationships between securities and market indicators
• Analysing alternative data such as weather forecasts and container ship movements
• Monitoring for suspicious transactions and triggering response protocols
• Generating reporting for clients such as portfolio and risk commentary
• Using chatbots and machine learning to answer employee or investor queries

All this points to one conclusion: in investing, AI is not a substitute for human skills and knowledge, but rather another, more sophisticated tool. Investment managers can add it to the toolbox they use in their quest for the best outcomes for their clients, while individual investors can use it to get access to data and trends they would otherwise not have had access to. Ultimately, though, the investment decision should rest with the human.

The value and income from a fund's assets will go down as well as up. This will cause the value of your investment to fall as well as rise and you may get back less than you originally invested.

The views expressed in this document should not be taken as a recommendation, advice or forecast.

^{1 https://resources.nutmeg.com/Our-investment-styles-explained.pdf}

^{2 Artificial intelligence | The next frontier for investment management firms}

This article was written by a human.