Artificial Intelligence 101 for Asset Managers

How will the industry be "disrupted" by AI? We break down the use cases for artificial intelligence in investment management.

A headline in Newsweek grabbed my attention recently: “Is North Korea a Threat? Not as Much as Artifical Intelligence, According to Elon Musk.” The article was in response to Elon’s tweet to the right.

If you’re familiar with Elon, then you are probably aware that this is no casual remark. He has put his money where his mouth is launching OpenAI, a billion-dollar nonprofit company, to work for safer artificial intelligence. When one of the world’s most lauded and visionary tech figures warns that “we are summoning the demon,” it’s perhaps wise to take heed.

Juxtaposed with the near-daily headlines touting AI in asset management, I’ve been thinking about whether our industry is headed down its own road to self-destruction. With passive management continuing to gain favor and sophisticated algorithms supplanting active management, are we writing ourselves out of the investment narrative?

AI promises to revolutionize the game of investment strategy by learning to predict outcomes. But when the world economy is impacted by the whims of humans (read: politicians, voters, leaders, media), I’m not convinced technology can forecast the fickle nature of these market factors any better than active managers can.

That said, there are some forms of technology that are extremely accurate when assessing outcomes and automating processes in investment management. As is the case with most new technologies, there is a spectrum of applications and possibilities. If you’re new to AI for asset managers, I offer a few definitions to delineate between the most discussed applications:

Robotic Process Automation—RPA is not the robot you’re picturing. These are simply algorithms that take an input and produce an output, as they are programmed to do. RPA can make rote middle and back office tasks extremely efficient.

Artificial Intelligence—AI is far more sophisticated than RPA in that it can perform calculations and actions that replicate human behavior. Robo-advisors fall into this category—they take unstructured and structured data in tandem and apply complex, pre-programmed logic to inform investment strategy.

Machine Learning—As AI has matured, technologists have explored how to program applications to make higher level decisions and store information to inform and refine future outputs. An example would be a system designed to do media sentiment analysis on Trump’s latest bill to determine its likelihood of being passed…then use that outcome to improve decision-making related to his next bill. In essence, this form of artificial intelligence programs a system to program itself.

Where do these technologies fit into the current fintech landscape? RPA is being actively implemented now and has the potential to disrupt middle and back office offshoring practices as algorithms replace the most rote tasks. For the latest, check out RPA leader Blue Prism to see how they are working with investment managers. Artificial intelligence is also in play with some investment managers and service providers, especially those who have a mature data governance structure. In September, State Street launched a project called MediaStats which uses "augmented" intelligence to read unstructured data collected from news articles, media reports and other textual information. Others are following suit. These technologies have the potential to disrupt traditional investment strategy and, at the very least, offer advantage in a competitive market.

As promising as these technologies are, I don’t believe AI or its subset, machine learning, will ever replace the active investment manager. If you’re of the same mind, the best way to reap the benefits of these technologies is to get your data house in order to establish the foundation needed to feed AI programs. And if you fear the worst like Elon? Perhaps you’re better off preparing the bunker.