This is the sixth and final report in Teneo’s ’23 & ESG Series.
Other reports in the series include the top 10 takeaways and key statistics of 2023 U.S. sustainability reports, DE&I data within U.S. sustainability reports, a study of Australian sustainability reports, sustainability report design considerations for 2024 and a European perspective on sustainability reporting.
For our final piece of the 2023 ESG Series, we examine one of the most impactful technological developments in decades: the business adoption of artificial intelligence (AI). In this piece, we highlight three important considerations for companies as to how AI may impact their sustainability reports in the future and what they can do to help prepare.
AI has quickly become one of the most talked about business topics in 2023. But before we get too far into AI, let us briefly define it. AI is the development of computer systems to perform tasks that normally require human intelligence. While generative AI – driven by powerful LLMs (large language models) grabbed headlines in 2023 for its ability to generate content ranging from imagery to music – more specialized types of AI are playing critical roles in less visible ways. For example, natural language processing (NLP) is a branch of AI that gives computers the ability to understand the meaning of and respond to data. Machine learning (ML) is a branch of AI that focuses on the use of algorithms to imitate human learning. For the purposes of this piece, we will refer to these and other branches of artificial intelligence collectively as “AI.”
1: AI is Already Being Used to Assess Company Sustainability Disclosures
Investors and their service providers such as ESG raters, proxy advisory firms and sell-side analysts are already using machines to assess companies’ ESG disclosures. It would be impossible for investors and others to assess tens of thousands of companies across the globe without it. For example, ESG raters currently use AI to read company sustainability reports as part of their ESG ratings analysis. As those tools become more efficient, ESG raters will likely evolve to a more “on-demand” rating system that can capture the impacts of world events on a company’s ESG risk profile in real-time. And just as investors already use AI to parse the nuances of a company’s quarterly earnings calls, investors are using AI to extract company ESG data, estimate company ESG data that is not disclosed, generate summaries to assess a company’s ESG goals and even identify red flags for greenwashing. For example, the University of Zurich and the Oxford Sustainable Finance Group have proposed a framework to assess the integrity of net zero plans. And this is just the early use cases for how investors and their service providers will utilize AI to assess a company’s ESG performance.
2. Companies Should Consider AI as One of its Key Stakeholders
With AI becoming the first stop in the ESG analysis process for many investors and service providers, companies should evolve how they think about the technology. Most importantly, companies should be using these tools to see what investors are seeing. While AI is, and will be an efficiency tool, it can also be fraught with errors, emissions or worse – entirely fabricated content (or “hallucinations” in AI lingo). Knowing what AI is saying about a company’s ESG strategy can inform sustainability reports, proxy statements earnings call transcripts and other communications. For example, companies should ensure that any graphics and charts are machine readable so that ESG ratings firms and other ESG data aggregators can ingest the information properly. In addition, language should be succinct so that AI can easily understand it and should not send mixed signals as that increases the likelihood of AI interpreting a company’s ESG communications unfavorably. Running a company’s (and their peers) sustainability reports through AI tools can help ensure that the ESG narrative will be understood in the intended way – whether the reader is a human or a machine.
3. Utilizing AI May Impact a Company’s ESG Goals
Using AI is an energy intensive activity. If companies embrace the use of AI in its business, it may have an unexpected impact on a company’s environmental goals. Whether it is the consumption of fresh water needed to cool the GPU “brains” of generative AI, or the 626,000 pounds of CO2 generated by training a single LLM, businesses need to account for the environmental impact of AI use. Company GHG emissions targets that were set before its use of AI may also need to be revisited for feasibility.
Companies using AI may also achieve greater efficiency, potentially leading to a reduction in the workforce. Algorithm bias is another critical issue that can meaningfully affect companies’ talent management from hiring to promoting processes. How do these potential issues square with many companies’ focus on their employees? Shareholder proposals on this topic have already been appearing on ballots, seeking to align the use of AI with employee overall well-being. Content management is also an area where AI based bias has been raised as a concern by stakeholders, as companies struggle with balancing user rights and the impact of uploaded content on society.
Using AI also creates opportunities for companies. In addition to enterprise models being offered by platforms like OpenAI’s ChatGPT, other third-party systems, as well as in-house models are being developed to synthesize company ESG data in a more efficient way than the traditional database/spreadsheet model of data storage. Boards of directors, who are typically responsible for the oversight of a company’s ESG strategy, may benefit from the use of AI for regular ESG reports. But questions remain about AI ESG data collection such as whether it is as accurate as traditional ESG data collection methods and whether an external auditor can verify it.
Looking Ahead
While 2023 was a transformational year for AI, we have not yet scratched the surface on how AI will impact businesses in the near-term. And we are still waiting on the dominant emergence of quantum computing which is set to be widely available by 2030. But it would be prudent for companies to at least begin to understand how AI is being used by investors and other stakeholders to assess a company’s ESG performance. Companies can use these AI tools themselves to give them a better sense of how their ESG communications are being assessed by stakeholders. In addition, maintaining close relationships with key stakeholders will also be critical to understanding the evolving landscape of how AI is being used.
