Wealth managers have a lot of data at their fingertip these days. They can precisely tailor investments to their investors' appetite for risk, using AI financial tools to gauge the business climate more accurately than ever before. But as a recent Financial Times article points out, AI is at a much more primitive state when it comes to ethical investing.
The problem isn't a lack of interest. Investors want to make investments that align with their values. The demand for environmental, social and corporate governance (ESG) investing is only likely to grow as consumers look to use their money as a tool for positive change. But the complexity of the data and the variability of consumer values present a complex problem — a problem that AI is just starting to solve.
The Challenges For ESG AI
Risks Vs. Values
Part of the difficulty is that calculating risk and calculating corporate responsibility are very different kinds of challenges. With financial risk, there are generally more or less objective factors to look at. AI models can derive reliable insight from financial statements, market swings, executive staff changes, and other factors that aren't particularly liable to spin. Even less objective factors like press coverage and public sentiment are beneficial as an aggregate, providing insight into investor behavior changes that you can use to stay ahead of the market.
With ESG, however, the information can be much more fragmentary, and the biggest source of info is almost always the company itself. No company is going to say they're doing unsavory things like recklessly polluting the environment, extracting minerals with forced labor, or allowing workplace harassment. There will be news exposes and third-party reports from time to time, but there will also be many puff pieces building up a company's purported social responsibility based merely on their say-so. Coming up with models to balance a heavily biased dataset towards corporate PR isn't impossible, but it is a difficult problem.
Calculating for Investor Values
Another issue is the range of investor values. Investors have different priorities, politics, and appetites for social and environmental impact. Some investors will want to fill their portfolios with the most environmentally responsible companies they can find. Others will focus on workers' rights, minority ownership, political affiliation or other factors. You'll also likely have customers with objections to particular industries, such as corrections or arms manufacture.
The intricate web of business relationships further complicates this. Companies can employ complex networks of brands and affiliations to legally segment different divisions that are part of the same enterprise or work extremely closely with companies whose practices an investor might object to.
The Future of Values Investing AI
All of these problems are difficult, but none of them are impossible to solve. AI can and will become an indispensable aid for value investing in the near future. And as in other areas of AI, value investing models will progress rapidly, providing better and better insights for investors.
The key to overcoming the challenges will be focusing on specific insights rather than an overall ranking. The gaps in ESG information and ubiquity of corporate spin make it unlikely anyone will develop a reliable overall ESG ranking soon. The range of investor values means that a unified ESG score wouldn't be that useful to many investors even if one could be accurately calculated.
Success will come from looking at the specific information available — even when it is partial — and providing details that consumers can evaluate against their values. By looking at a wide range of sources, including news and third party reports, we can learn a lot about how a company treats the environment and communities it operates in and empower investors to make more ethical decisions.
To learn more about how AI can keep investors informed, check out our white paper on how Bitvore was able to use news datasets to locate Amazon's HQ.
