EdenTree Investment Management's responsible investment analyst, Cordelia Dower-Tylee, has shared insights on the burgeoning role of artificial intelligence (AI) in evaluating environmental, social, and governance (ESG) risks in investment portfolios. With the growing demand to integrate ESG considerations into investment research, AI has emerged as a revolutionary tool offering novel methods for data collection, processing, and financial analysis. However, Dower-Tylee warns that over-reliance on AI could lead to a skewed understanding of a company’s real impact if not applied meticulously.
AI's potential in ESG analysis lies in its computational capability, which offers significant benefits in overcoming challenges within the sector. Analysts are expected to gather vast amounts of data from varied sources such as corporate reports, news articles, and scholarly research, to perform comprehensive ESG assessments. AI aids in efficiently collecting this non-linear data, allowing analysts to allocate more time to interpreting the information, which is a more effective use of their expertise.
Moreover, AI can assist in standardising ESG data, which is often inconsistent and unaudited, unlike financial reports. This capability facilitates easier comparisons across companies and industries. AI can also fill in gaps by estimating data and enhancing consistency between datasets. For instance, AI provides valuable insights into site-based biodiversity indicators where data scarcity and risks are prevalent.
Sentiment analysis, a technique that evaluates the emotional tone of documents, exemplifies AI's capacity to deliver new insights. This method can reveal attitudes towards initiatives like low-carbon transitions by assessing tonal trends in corporate disclosures over time.
Despite these advantages, Dower-Tylee advises caution against depending solely on AI for managing ESG portfolios, highlighting potential for misrepresentation. AI outputs often render binary results devoid of context, overlooking the nuanced dimensions of ESG issues. A key concern is the difficulty in quantifying aspects like employee wellbeing and corporate culture, which are not as straightforward as carbon emission metrics. AI might miss these qualitative subtleties, leading to erroneous conclusions about risk management practices.
Human intervention is essential to interpret AI-generated outputs and correct inaccuracies, particularly given the complexities in aligning companies with their actual parent entities. Without such supervision, there is a risk of inflation or diminution in assessed risks or impacts. Moreover, the nuances of ESG elements cannot be fully appreciated in isolation; they exist within broader contexts that AI may miss. A pertinent illustration is how geopolitical and environmental events have influenced companies like Enel in their emissions targets, where a human perspective provided a more comprehensive understanding.
Additionally, human engagement plays a vital role in ESG analysis, offering insights through direct interaction that AI cannot replicate. This method provides a deeper understanding of a company's culture and objectives, reinforcing investment recommendations.
While AI is invaluable to ESG analysis for its prowess in data organisation and pattern recognition, its role should complement, rather than dominate, human-centred analyses. As ESG analysis inherently involves complex and layered issues, a nuanced examination by humans remains crucial. AI can enrich this process, but the core of understanding ESG issues remains a fundamentally human endeavour.
Source: Noah Wire Services