Big data

Intern Naomi Dobey also contributed to this column.

Over this past year, we have seen Nvidia and the rest of the Magnificent Seven (Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla) take over Wall Street. With Tesla’s artificial-intelligence-powered battery management, Microsoft’s conversational speech recognition and Apple’s seamless integration of AI in its most recent operating system, exposure to AI has become central in a well-diversified technology portfolio.

ETFs offer two routes to investing in AI: finding ETFs that have exposure to companies such as the Magnificent Seven, or investing in ETFs with models that have an AI-based stock-selection process.

Regardless of the route investors take, the risks must not be ignored. AI models base themselves on a variety of information from diverse sources (verified or not) to create an investment framework. The source of the information can be subject to manipulation and human error, even from experts.

This could result in decisions that would be very different from what a human would make. This is why AI is not necessarily ready to make investment decisions yet or adapted for industries where several nuances are linked to successful outcomes (for example, health).

Furthermore, companies like Facebook and OpenAI often rely on low-paid labour to handle data labelling. Poor data will result in a poor model, and relying on such models for building an investment portfolio can come with risks.

It is also important to make the distinction between automation and artificial intelligence. While automation is rules-based, AI is based on intelligent algorithms that can recognize patterns or features and learn from those. Beta or factor ETFs, which are rules-based, have been on the market for more than a decade, but AI-based portfolio-model ETFs only date back six years. The AI Powered Equity ETF (NYSE Arca: AIEQ) launched in 2017 and uses IBM’s Watson for portfolio analysis and stock selection.

Not all AI-themed ETFs have survived the test of time, with some having been liquidated or delisted. These include the Horizons Active AI Global Equity ETF, Emerge ARK AI & Big Data ETF and EquBot AI Powered International Equity ETF. I’ve previously written about performing due diligence on new ETFs.

Nonetheless, in this growing area of investment, many ETF options remain available, such as the iShares Exponential Technologies ETF (Nasdaq: XT) and the Global X Artificial Intelligence & Technology ETF (Nasdaq: AIQ). If you wish to venture into the world of AI-based modelled ETFs, you might consider AIEQ and the Qraft AI-Enhanced U.S. Large Cap Momentum ETF (NYSE: AMOM).

For now, however, I am not recommending any specific ETFs. I continue to overweight technology in general using the Technology Select Sector SPDR Fund (NYSE Arca: XLK). It’s important that investors understand how an additional thematic ETF in the AI space can overlap with existing holdings.

Should AI-related ETFs be part of your investment portfolio? My guideline, as I’ve written about before, is that thematic positions should be 5%–10% of an overall investment portfolio.