The Emergence of AI Powered Funds

First appeared on fundsglobalasia.com / funds-europe.com

In the rapidly evolving world of asset management, artificial intelligence (AI) has captured the imagination of many investors. With its ability to consume large unstructured datasets and apply algorithms to generate analysis, AI appears to have strong potential to drive investment decisions for better returns.

One recent example is in Singapore, where the first AI-powered, actively managed exchange-traded fund (ETF) was listed on the Singapore Exchange (SGX) in January 2024. The fund leverages its proprietary AI and machine learning models to research the market and select stocks for investment. The commonly cited benefits of using AI includes its ability to process real-time market data and the avoidance of unconscious bias leading to more dynamic and rational investment decisions. 

With a handful of AI-powered ETFs entering the global market, the discourse revolves around whether these algorithms can indeed deliver on a promise to deliver superior returns. 

Notable examples of AI-powered ETFs with a multi-year track record include the AI-Powered US Equity ETF (AIEQ), launched by Equbot in October 2017. AIEQ employs a sophisticated AI system, leveraging natural language processing and machine learning algorithms, to analyse vast datasets and identify investment opportunities. Another example is the Qraft AI-Enhanced US Large Cap Momentum ETF (AMOM), launched in May 2019. Developed by Qraft Technologies, AMOM utilises AI algorithms to identify momentum trends in the US large-cap equity market. By dynamically adjusting its portfolio based on real-time market signals, AMOM aims to capture alpha while mitigating downside risk. 

Both AI-powered ETFs performed at parity with the S&P 500 in 2022, but the past 18 months paints a different picture.

While these examples showcase the ambition and potential of AI-powered funds to deliver superior returns, track records are mixed. Both AI-powered ETFs performed at parity with the S&P 500 in 2022, but the past 18 months paints a different picture. Since the beginning of 2023, AIEQ has trailed the S&P 500 significantly, underperforming the popular benchmark index despite the backing of IBM Watson and quantum computing. While AMOM kept pace with the S&P 500, the promise of combining human intuition oversight and the analytical depth of AI did not result in a breakaway success.

Market 'noise' impedes AI

One of the challenges for AI in making investments is the contradictory market sentiments from a broad spectrum of commentators, ranging from professional market analysts to individual market watchers on social media. According to the OECD report on Artificial Intelligence, the varied data sources made it difficult for the AI to decipher the authenticity of the information. By consuming all the “noise”, it impedes the AI from making more accurate judgements, as alluded by the concept of “garbage in, garbage out”.

Algorithmic hallucinations and the generation of false or misleading outcomes, compounded by AI’s inability to accurately point to the sources of its recommendation, further inhibits the uptake of this technology in asset management. To professional investors, one erroneous prediction can lead to millions in investment losses and reputational risk for asset managers. 

While the advent of AI enabled asset managers is exciting, it is essential to approach this paradigm shift with caution and prudence. The lessons we have learned from the small initial wave of products, such as the above-mentioned funds and the additional risk exposure from incorporating AI, highlight the long journey ahead to fully capture AI's promise and the importance of robust risk management and human oversight. 

As informed professionals and asset owners navigate this changing landscape, they must strike a delicate balance between harnessing AI's potential and preserving the integrity of human judgment and oversight. In the dynamic interplay between man and machine, the future of asset management lies in embracing AI as a complementary tool to augment, rather than replace, decades of investment experience accumulated by investment professionals.