Artificial intelligence (AI) continues to build its presence across global industries, so its progression into the asset management sector is only natural. But AI can be understood differently by different people, so definitions are important.
One should think of AI as a field in computer science that builds “intelligence” into electronic systems. These AI systems are able to perceive data environments, learn, and take actions to achieve objectives. This definition presents a compelling value proposition for AI versus traditional portfolio management systems with respect to today’s growing data environment. The key technological advantage of AI investment platforms is their overall flexibility to appropriately process massive amounts of dynamic market data and identify investing opportunities.
AI can help us better understand what to trade, and how to trade it. Important AI capabilities include: consuming, learning, and aggregating growing volumes of data across mediums, continually adjusting portfolio risk based on observed market signals (and the concurrent removal of rigid factor-based criteria), and the ability to connect new market signals and derive an optimized portfolio free of human bias.
High-performing systems should be able to assimilate and manage both structured and unstructured data in a timely manner, and appropriately process erroneous or blatantly fake financial news. Many early stories of AI investment systems described the brute force design of jamming massive data sets into the conceptual “black boxes” and allowing the machine to produce a series of recommendations. Better solutions exist and are more suitable for the explosion of available investment data and the demand for system operational observability.
These more transparent and higher-performing AI data processing systems are likely to drive increased asset flows into the space. As this evolves, investors should recognize that not all AI is created equal.
In reviewing the emerging class of AI-powered funds, an effort should be made to understand how they process data — the same kind of due diligence that would be done on a human portfolio manager. The truth is that AI investment platform design will vary significantly, and in turn influence performance. But it is also true that the strongest AI-powered approaches coming to market do not operate as “black boxes” but instead are overseen by teams that should be able to clearly articulate their respective approaches. They should also explain just what drives their algorithms and the types of opportunities that have been engineered to uncover. […]