The Risks and Opportunities of AI

Artificial intelligence (AI) is evolving at an astounding clip, expanding the frontiers of technology, and affecting how businesses evolve and how we, as investors, evaluate opportunities. We are still in the early days of real-world deployment of advanced AI applications, but it is estimated that annual AI investments will quickly reach into the hundreds of billions of dollars and that AI has the potential to boost global productivity meaningfully, perhaps by more than one percentage point per year in developed economies.¹

At Boston Trust Walden, we have been examining AI across multiple dimensions, including its impact on society, individual companies, and how we operate our own business. While we suspect the full impact will not be known for years, the ramp-up in spending is occurring now. It is also increasingly apparent that there are pitfalls and risks associated with the latest technology, in addition to likely opportunities. As we try to better understand the implications of this evolving technology on our investments and our company, our investment philosophy remains unchanged; we continue to seek high quality, sustainable companies whose stocks are trading at reasonable valuations.

AI Past & Present

The term artificial intelligence encompasses a wide range of technologies. The more basic of these, such as certain machine learning algorithms that have been in wide use for years or decades, have already made a significant impact on many industries. From financial modeling to pharmaceutical development to IT security, the influence of machine learning models is profound.

More advanced and sophisticated AI technologies, such as generative AI built upon large language models utilizing neural networks, have been in existence for several years but have only recently stepped into the mainstream. While these newer models have yet to trigger transformational change, it’s only a matter of time before they mature to a level where companies can rely on their output for their production processes, and more tangible use cases emerge.

Sophisticated AI has the potential to be disruptive, but disruption is not necessarily bad. In some ways, AI will be good for society, such as enabling greater productivity in the macroeconomy. It will also have negative effects; for example, workers may be displaced when jobs are lost. The Hollywood writers’ strike in 2022 was, in part, organized to address such a threat to writers’ jobs.² Like past leaps in technological progress, we can be sure things will change, but cannot be sure where the changes will lead. And as with past examples of rapid technological change, there will likely be “winners” and “losers.”

Investment Opportunities & Threats

With the meteoric rise of AI, clients have naturally begun to ask us where we are finding opportunities to invest. Broadening adoption of powerful AI tools will propel growth for some companies, while others will find their business models at risk without a natural avenue to replace lost revenue streams with new ones. At present, the companies that are at the forefront of developing these technologies or providing the necessary tools for their development are the ones most impacted by the rise of this newer technology. These include both start-up and established companies that are currently in a race to build the most sophisticated AI models, as well as semiconductor manufacturers and other firms that provide the considerable computing power required for their construction and use. The list of AI “winners” will likely broaden in the future to those companies that are successful at leveraging the technology to provide new and proprietary services built upon it. Companies involved in text or image editing are already rolling out such offerings, and more is certainly to come.

But while some companies will be winners, gaining market share and growing profits, others will lose out. Companies that cannot compete with the technological improvements of their peers, either due to lack of vision or lack of resources, may lose share. Likewise, certain services may become obsolete as consumption patterns shift to incorporate new advancements in technology. This trend has plenty of precedent, from buggywhips to landline phones to video rental stores, and it will no doubt continue with advancements in AI.

At Boston Trust Walden, we aim to identify companies with sustainable business models, including a mechanism to sustain past success into the future. To do this, we consider the company’s existing sources of competitive advantage to determine its future prospects. However, with AI’s potential to disrupt business models, we must consider how competitive dynamics may play out, and whether current advantages can be defended.

Interwoven with the financial considerations of AI-related effects on companies’ business models are the related environmental, social, and governance (ESG) risks and opportunities. The writers’ strike is an early example of a group bringing “license-to-operate” issues to the fore, at least within the context of one industry. There are also significant environmental concerns given the increase in computing power — and associated cooling of AI servers — needed by these models relative to more primitive computer models or internet search. For companies using AI, data security and privacy, corporate governance, human capital, and regulatory risks will need to be navigated carefully as the AI landscape evolves and will thus remain important components in our evaluation of the attractiveness of stocks and their inclusion in client portfolios.

As we navigate the opportunities and risks to companies, we continue to also focus on their valuations. Euphoria over growth prospects, or fear of challenges, can drive stock prices to unreasonable valuations. While we evaluate which high quality companies may be AI winners, we also keep a close eye on valuations so that we don’t pay more for a company than its realistic growth expectations support.

Our Approach to Using AI

For investment research at Boston Trust Walden, AI is a welcome addition to our toolbox. Aspects of our quantitative analysis already incorporate some machine learning models. As newer AI technologies like large language models mature, we will continue to evaluate how best to incorporate them into our existing decision-making process. However, our investment philosophy centered on high quality, sustainable companies trading at reasonable valuations will remain fundamentally unchanged, with final decisions being made by portfolio managers.