The following is a guest article from Cyrus Mewawalla, Founder of CM Research, which is the first installment of a two-part examination of the impact of artificial intelligence technology on the investment management community—and investment research .
AI is the path to maximum profitability
Artificial intelligence (AI) threatens to disrupt every industry – from heavy manufacturing, to retail to investment management.
AI refers to software-based systems that use data inputs to make decisions on their own. AI software typically has three attributes: it can recognise sound, images, videos or natural language, enabling it to make sense of data inputs that traditional computer systems cannot understand; it can find answers to questions by forming fields of hypotheses; and it can learn from experience.
The key technologies behind AI are machine learning coupled with natural user interfaces such as voice recognition, image recognition, gesture control and context aware computing.
Leaders in the field of AI include Google, Amazon, Facebook, Microsoft and IBM. A number of smaller players – Palantir, Kensho and Sentient – are particularly focused on disrupting financial services industries, including asset management. Their aim is to dominate the information flow and analytics section of the value chain, leaving the boring mechanical bits to the industry incumbents. Therein lies the path to maximum profitability.
AI will change the investment landscape in three ways
Artificial Intelligence will change the investment landscape in three ways. First, investors will be forced to place more value on the quality of a company’s AI assets. Second, investors themselves will rely far more on AI-based research techniques to support their investment approach. And third, investors will compete head to head for AI talent with the technology sector.
Let us consider each of these in turn.
The advance of complex, scientific algorithms – increasingly in the form of algorithms that learn from experience – is causing anxiety in boardrooms across the world.
Self-teaching algorithms already deconstruct and predict patterns in everything from human behaviour and asset prices to the operation of jet engines and autonomous cars.
These algorithms, coupled with the proprietary datasets that they feed on, form the basis for the disruptive power that has yet to be unleashed by the big internet ecosystems such as Google, Amazon, Facebook and Baidu. They are also the drivers that are transforming GE from a lumbering industrial megalith into a data driven industrial applications company and IBM from a legacy IT hardware manufacturer to a cognitive software giant.
Algorithms are eating the world. Whether it is Google search, Amazon Recommend, Uber Surge, Facebook News or Google’s AlphaGo, more and more of a company’s competitive advantage now comes in the form of some kind of algorithm.
In March 2016, the AlphaGo programme developed by Google’s DeepMind division scored a high profile 4-1 victory over the world Go champion. This clear-cut victory hinted at what deep learning algorithms might be capable of down the line with enough computing power and data behind them.
For many, the real surprise was that the Google team had achieved a level of artificial intelligence that many industry commentators believed to be at least a decade away.
Meanwhile, the core theme at last October’s Gartner Symposium ITxpo was “The Algorithmic Economy”. At this global IT conference, thousands of computer industry executives were told by Peter Sondergaard, Gartner’s Head of Research, that the algorithmic economy “will power the next great leap in machine-to-machine evolution in the IoT… products and services will defined by the sophistication of their algorithms and services…”
Gartner expects the Internet of Things to swell to over 20 billion “sense aware”, connected “things” by 2020. These things will include smart phones, eye lenses, clothes, connected cars, car tyres and construction sites.
AI is the next big technology platform, packaging machine learning with natural user interfaces (such as voice recognition or gesture control) to create products that can literally obliterate entire industries or professions.
AI is the revolution that is brewing over the next five years. Its transformational impact is on par with that of the mobile phone or, before that, the PC. But for AI to work well, it requires data to feed on – and lots of it.
This returns us to familiar network effects and feedback loops that characterise the Big Data economy and the economic logic of returns to scale: the more people that use the systems, the smarter the systems become; the smarter the systems become, the more people who use them; and so on.
It is why, from the earliest days, Larry Page used to stress that Google had not created a search engine but “an AI” and it is why, having rolled up so many users, the big internet ecosystems (Apple, Amazon, Facebook Baidu, Alibaba, Tencent and others) have built broad and deep moats around themselves.
This is the world that professional investors are having to get to grips with. When valuing a company, equity analysts like us will soon start to ask the question, “Is this company an AI powerhouse?” or “Is this company dangerously exposed to AI?”
Indeed, we at CM Research have already begun the arduous process of incorporating the AI capabilities of the 500 TMT companies we follow into our screening process. We predict that, by 2025, 20% to 30% of the top 1,000 global corporations will be eviscerated by their failure to assess the risk of AI-based threats to their core business.
A company’s earnings prospects will increasingly be defined by the extent to which it incorporates AI systems within its products.
The next installment will explore how investors will rely more on AI in their research process, the war for AI talent and the timing of AI.