Three Thoughts on Artificial Intelligence in 2021

2020 could be said to be the year of AI everything, when AI applications seem to be entering into every sphere of daily life. With investors being presented with strong evidence that AI strategies are the new alternative investment, here are our three thoughts on AI in 2021.

1. AI Everything

There is building investor confidence that with AI a new page in technological progress has been turned and with it a raft of new business opportunities will present themselves. The field of AI in asset management is in a particularly strong position to gain in 2021. With many investors seeking yield in the post pandemic environment and some strategy types delivering unexpectedly disappointing results in 2020, AI strategies pose a fresh alternative investment. These approaches are particularly attractive for those looking for liquid, easily benchmarkable investments.

Investors have accompanied AI on its journey over the past decade, with many of course, choosing to sit back and observe where things would go, before committing to invest in it. The industry has come a long way from the false dawn of a decade ago, when there was much speculative interest in AI but little tangible fruit to show for it.

Everyone recognizes that the technology has advanced significantly in the last ten years but less obviously, and possibly more importantly, there has been enormous progress on the development of different AI approaches. In 2021 we will likely see an increase in the use of nuanced AI approaches, such as in task-specific AI technology and increased hybrid usage, where AI is used in conjunction with and to complement human activities.

This evolution has seen AI become a recognizable presence in much of the world around us as it is integrated into and used to enhance familiar systems across many sectors, with applications as diverse as the assessment of oil wells to the analysis of medical imaging.

2. The Translucent Box

Investors are familiar with ‘black-box mentality’ which is synonymous with systematic trading. As a consequence of the complexity of AI systems, there is a strong onus on AI managers to be able to explain the investment process. This helps investors build trust in AI strategies. It takes them out of the black-box, the unknown, and enables them to be understood from a perspective familiar to the investor, without expecting them to have full grasp of a complex technology.

There is a greater demand emerging for AI systems to be humanly explicable, particularly when they are not easily comprehensible. This is a very useful development as it forces a clarity on developers of AI systems. This in turn helps separate ‘real’ results (those which are explicable) from those which are simply ‘correct’. This is a strange facet of AI development: You can end up with correct results and yet have no idea how the system obtained them. In some fields of deployment where the results being obtained are not of much consequence, how they were obtained is not so important. In other areas, particularly those which demand a sensitivity to human beings, however, being able to explain the how and why of it, is crucial to prevent people feeling alienated from the technology and fearful of it. Asset management fits into this latter category.

3. Regulation and Data

For participants in the financial sector regulation is so much part of the furniture that we expect it to be there and when it is not, we attempt to gauge when and in what form it will materialize in the future. Regulation is an anathema for the technology and communications sector and is treated like a sudden inconvenience with the potential to become its nemesis. AI fits very much into the tech mold and many of its users have casually assumed that they will have free reign on data and information usage without any interference or interruption. This fails to recognize the degree of change that AI is bringing to our daily lives. The extent of this change, and peoples’ sensitivity to it, is such that regulation is inevitable.

The torchbearers of regulation, the EU, provided a recent illustration. The EU Agency for Fundamental Rights issued a warning on the use of AI with respect to the rights of the individual in a report released on December 13th, forming part the process to establish EU rules on AI usage.

This report was focused on issues pertaining to security (with the potential for government overreach), medical diagnosis and targeted advertising, concerns regarding personal data/information rights. It can, however, be viewed in a broader context. The gradual definitions being imposed on who owns, or can use, what data, has been an elephant in the room when it comes to over reliance on big data. The tendency to simply pump more and more data into AI decision-making systems will start to get tripped up when it comes to suddenly being unable to use some data set or data subset due to regulatory or ownership constraints. This prospect has the ability to derail many overly-complex decision-making processes which can be exposed to vast arrays of data. 2021 will likely see the slow-moving wheels of regulation churn toward AI. Hence it would be advisable to seek out AI opportunities which have already taken a proactive approach with regards to how they insulate their processes from data-usage dangers. One example would be those with a focus on using derived data.

© 2020 Plotinus Asset Management. All rights reserved.
Unauthorized use and/or duplication of any material on this site without written permission is prohibited.

Image Credit: Kentoh at Can Stock Photo Inc.