Black Box Investing 2.0

This past week, it was announced that a San Francisco-based tech firm, EquBot was launching the first ETF to use artificial intelligence (AI). EquBot is built on IBM Watson, a platform which rose to prominence from its highly publicized success as a robot contestant on Jeopardy. The ETF applies proprietary algorithms to Watson’s AI technology to build predictive financial models on 6,000 publicly traded U.S. companies. It then ranks the companies in terms of risk and return potential to create concentrated portfolios of 30-70 holdings. The goal is to beat the broad market indices without incurring additional risk.

In recent years, a number of investment firms including Sentient Technology, Numeri and Emma have employed AI in various ways to improve returns. But these pioneers, have to date, been hedge funds or private investment startups so there has been limited information on performance results and limited access to the products for retail investors. EquBot will be the first opportunity for the public to weigh the performance results of a true AI engine and be able to include it in their portfolios.

It will be interesting to see if EquBot can post consistently above market returns out of the box. It will be equally, if not more, fascinating to discover whether it can improve its performance over time as the machine continues to learn and adapt.

Material performance success by EquBot is likely to lead to a significant rise in AI managed products going forward, perhaps even creating a broad new class of investments. It may at the same time reawaken the argument around the efficiency of markets and the relative value of passive investing.