AI and responsible investing: four things you should consider

AI and investing is already a hot topic. Impact investing and ESG - is also a hot topic. However, they're not being spoken about at a great deal together. Don Gerritsen, Head of Benelux and Siobhan Archer, Associate Signatory Relations, UNPRI aim to change that.

The use of Artificial Intelligence (AI) is rapidly gaining momentum within the global investment community, including the PRI’s signatory base.

In 2018 year, the PRI brought out a technical primer on blockchain, skimming the surface of the disruptive new technology that investors had to hand.

However, as AI grows and we see signatories upgrading their RI practices, we’ve taken a step back to look at some of the way AI is making changes. Acknowledging some of the exciting possibilities AI presents to responsible investment and some of the challenges that signatories will face.

AI promises opportunities around four different areas: decision making, the workforce, interaction with clients and beneficiaries, and advanced use of data.

1. Advancing investment decision making

The link between neuroimaging and decision making has been studied for the last decade, with reward-related decision behaviour fascinating investment professionals. The analysis of investor performance data and their brain activity is one way AI can enhance decision making. An ambitious feat, it is something researchers at Bonn University have investigated in the way brain activity cautions against buying certain stocks.

Using AI algorithms, researchers were able to build a model that helps understand behaviour around stock selection and reward related behaviour. Although this is still in early stages, investors expect this to dramatically enhance decision making around picking investment opportunities through it’s ability to recognise patterns.

According to statistics from Accenture, it’s estimated that through the use of AI, global GDP can be boosted to $15 trillion by 2030.

And yet, although AI can enhance decision making in the investment process, it is important to keep in mind that some foundations of good governance, like ethics; intuition; emotion and common sense, are inherent to humans. AI controls for human error, however this raises questions around how important emotions and gut feelings are to investors?

Previously traits that were only attributed to humans, Harvard Business Review suggests that AI is developing emotions, in 2009 Philips and a ABN AMRO discussed tracking investors’ moods through a bracelet, warning them not to make trades when emotional. The tracker works on pulse detection responding with a red light to warn investors to reconsider before acting irrationally. Identifying these trends may help investors to encourage more responsible behaviours especially around E, S and G factors.

2. Workforce

In addition to improving the decision-making of the human workforce, AI also has the potential to raise productivity. In many industries, including finance, it’s already replacing human labour.

According to statistics from Accenture, it’s estimated that through the use of AI, global GDP can be boosted to $15 trillion by 2030. In the United Kingdom alone, a 25 percent increase in labour productivity has been forecast by 2035.

However, with the benefits, also come significant social and transitional risks with the potential of some 30% of jobs that are at potential risk of automation by the mid 2030s.

China is a telling example: 23% of jobs in the country's financial sector will be cut by AI or will be transformed by 2027 as predicted by BCG.

Those jobs that are the most at risk are those that can be easily replaced by machines, often jobs that are given to low-skill workers. This raises questions how investors can leverage a smooth transition to a low carbon economy and seriously consider the risk of stranded assets as well as costs around unemployment and retraining staff to do other jobs.

Responsible investors will be aware of how workers at all stages in supply chains will be affected by AI, and should be aware of how social risks can be integrated in to the capital allocation across asset classes.

3. Anticipating clients’ and beneficiaries’ needs

Big data and AI provides prospects for more personal servicing of customers and beneficiaries by anticipating needs and identifying characteristics of high-value customers.

A smarter and more targeted client outreach through AI, enables advisers to understand investor preferences and better manage and tailor content for specific clients. Examples include investors using alternative data sources such as engagement on social media, improving the way in which products are distributed. In this way, client’s needs can be met faster providing a better customer service experience.

While a number of questions need to be answered, AI provides opportunities for investors to advance decision making and efficiency.

One such example is APG's use of algorithms to enhance personal interaction when servicing clients. The firm, which also employs data analytics for sustainable investment, uses a number of data points including contact history, personal data, pension details, and the online browsing behaviour of large participant groups to predict when and for what reason they will contact the pension fund ensuring their clientele’s needs are met.

This trend in AI has also been echoed in the recent rise of robo advisors especially with regard to retirement planning.

Nevertheless, there are cases that show that imperfect AI systems may generate data that discriminates against certain groups, such as facial recognition software conducting unintended ethical and gender profiling.  Data may also be exploited when used by non-authorised entities: Facebook and other media companies have been used as prime examples of how data privacy affects stock prices and consumer confidence.

4. Enhanced data

Artificial intelligence based algorithms have dramatically changed the way investors access alternative data sources. Advanced quant funds have been known to scrape the internet and public databases for relevant RI metrics. PRI signatories like Arabesque Asset Management review 65,000 data points including in-depth news analysis when selecting stocks, an accomplishment only available through AI systems.

Additionally, investors have starter using Natural Language Processing and semantic modelling to assess company performance based on qualitative web data and web-mood analysis with the addition of risk warnings to anticipate potential market shocks.

The accuracy of analysis by AI systems, however, remains debatable. The rise of fake news can lead to an increase in online inaccurate and contradicting data. The risk of having data so easily available to AI algorithms means that conflicting and contradicting data may be taken into account and risks skewing models. Another ESG risk that is promoted through the use of AI is the lack of accountability, removing the human element will make it more difficult to identify where problems originate and distancing corporates from mistakes.

While a number of questions need to be answered, AI provides opportunities for investors to advance decision making and efficiency. But especially when it comes to responsible investment, a balance between human and artificial skills must be struck.

Don Gerritsen will be speaking at FundForum Global ESG and Impact, 12 - 13 November 2019 at the Amsterdam Marriott, Amsterdam. He will speak on a pivotal year for business and sustainable finance.