Environmental, social and governance (ESG) and artificial intelligence (AI) are two of the biggest trends in the finance world, but working together they can potentially transform how businesses operate.
One of the biggest developments within the world of AI has been the proliferation of generative AI tools. Companies like OpenAI have shown how powerful generative AI can be, with ChatGPT and Sora both showing how easily a simple set of instructions can create some impressive results. The potential of generative AI has captured the interest of the financial sector. In fact, generative AI within the financial services industry is expected to grow at a 28.1% CAGR to reach a value of $9.4bn by 2032, according to Statista.
Internal ESG processes could be the natural home for generative AI. Not only are consumers increasingly becoming interested in ESG factors, but regulators are also increasing their attention to these factors. As a result, businesses find themselves needing to improve their internal workflows to ensure they meet new regulations, such as the EU’s new corporate sustainability reporting directive (CSRD).
Position Green managing director Brussels Julia Staunig said, “Generative AI can help users understand gathered data more deeply, see trends from seemingly unrelated data, guide them towards ensuring all relevant data has been gathered, and give actual recommendations on actions to try to reduce emissions based on a company’s specific situation amongst many other uses.”
IntellectAI’s CTO Deepak Dastrala also expressed confidence in the use of generative AI within ESG and outlined several potential enhancements to ESG processes. The first of these was finding and consolidating trusted ESG data. For example, generative AI could intelligently navigate and consolidate ESG data from diverse sources, addressing fragmentation within the ESG data ecosystem and building a unified and complete dataset.
Another use case would be for the creation of customised ESG insights. Deepak said, “This would allow for differentiated insights, or focusing on topics without good coverage such as oceans or biodiversity”
Other enhancements through generative AI would be to provide estimates and simulations of ESG metrics to fill data gaps, empowering in-house ESG experts to independently build datasets, fostering transparency and control, automating ESG reporting, and analysing unstructured data to help uncover insights and trends.
Improving ESG analysis
While generative AI could bring improvements across a range of ESG related tasks, one of the biggest areas of impact is with data analysis. Staunig noted, “Within analysis specifically, models like GPT-4 have the potential to look at an entire emissions dataset, and at the same time understand an entire organisation, to then for example help figuring out why and how your emissions data looks like it does: Deeper insights than before.”
Given the benefits generative AI can bring to ESG analysis, it comes as no surprise that Position Green has its own solution doing just this. The ESG FinTech company launched its AI Analyst feature in the tail end of 2023 in a bid to help firms transform their ESG reporting capabilities. The new tool, which is integrated with Position Green’s Sustainability Management Suite, leverages generative AI to transform data into actionable insights in real-time by detecting patterns and anomalies, as well as highlighting key findings to review.
Demonstrating the significant market size that generative AI can capture, IntellectAI also has its own generative AI tool. Intellect’s generative AI platform powers ESG Edge, which is used by sophisticated corporate investors to generate unique ESG insights across their investment portfolios. As a single question to your own guidelines, and get the relevant ESG intelligence in response to every single company in your portfolio.
Deepak noted a number of potential benefits firms could expect to see from leveraging generative AI for their ESG analysis. Among these, scalability is a chief benefit. Generative AI will enable firms to scale their analytics for diverse portfolios without needing proportional increases in cost or complexity.
Generative AI would also allow firms to identify unique sustainable investment opportunities, improve the precision of risk metrics for better-informed investment decisions and risk management strategies, predictive capabilities and pattern recognition to enable proactive ESG-related risk management.
Additional benefits include the ability for wealth managers to create sophisticated ESG analytics that are accessible to a wider range of firms and reduce the cost and time required for ESG data collection, analysis and reporting.
What to know before using the technology
A lot of hype has been built around the possibilities of generative AI, encouraging many firms to eagerly look at how they could potentially implement it. However, it is important to not be blindsided by the excitement and leap straight into integrating the technology. There are several factors to consider before choosing to adopt the technology.
An important aspect to be aware of is that AI is not infallible. As generative AI is an impressive technology, it can be easy to assume it will never be wrong, but this is not the case. AI hallucinations are occurrences that can be caused for a number of reasons, whether it is incorrect source data, gaps in the data or just a misunderstanding of prompts, there are a number of ways that an AI technology stack can lead to a wrong answer. Similarly, inherent biases in the training data can also lead to skewed ESG assessments and create severe ethical concerns. As a result, it is vital firms have strategies to put data output through the quality grinder, check for accuracy, relevance, toxicity, gaps in coverage, data drift and more.
Deepak added, “The complex nature of generative AI models can also pose challenges in terms of transparency and explainability, complicating auditability and accountability. There’s a risk of creating a dependency on AI for ESG analysis, potentially leading to a skill gap in traditional analysis methods among analysts. Additionally, the capacity for generative AI to produce convincing narratives could be exploited for market manipulation, while mishandling or misinterpretation of ESG data generated by AI could severely damage a firm’s reputation.”
When it comes to looking to implement the technology, Staunig said, “Use general and capable models first, to make sure the output is interesting, before going to more niche models. Make sure you have narrow-scope ideas to validate first.”
As for Deepak, he encouraged firms to establish strong data verification processes to ensure AI-generated data is accurate and reliable. “Prioritising transparency and auditability by investing in explainable AI technologies can help demystify the decision-making processes of AI models, building trust among stakeholders. Cultivating AI literacy and critical analysis skills within the organisation is also crucial, enabling teams to effectively interpret and validate AI-generated insights.” On top of this, selecting the correct data sources will be vital for ensuring the underlying data used by AI analysis will be relevant and high quality.
On a final note reflecting about the potential of generative AI within ESG, Deepak said, “The integration of generative AI into ESG practices represents a significant leap forward, with the power to redefine sustainability and responsible investing. Central to leveraging this transformative technology is the focus on Trusted Data. Ensuring accuracy, transparency, and ethical use of generative AI is crucial to building trust in its applications. By mitigating potential risks and adhering to best practices, firms can fully utilise generative AI to achieve impactful ESG outcomes.”
Keep up with all the latest FinTech news here.
Copyright © 2024 FinTech Global
