The ESG Risks of AI – and How to Manage Them
Chup Priovashini, Rory Sullivan & Robert Black
AI is changing how businesses operate, and with it, comes changes to their ESG risk.
AI adoption has increased dramatically in the past few years with the advent of widely-accessible generative AI. Over 70% of organisations derived value from AI in 2024. AI is also touted as a strategic growth driver by the UK government, with the potential to boost the economy and provide jobs.
While AI undoubtedly presents a huge opportunity for businesses, its widespread adoption also comes with risks which must be addressed by long-term investors.
Chronos Sustainability has worked closely with Railpen to develop its report Achieving effective AI governance. Alongside welcoming the significant potential benefits of AI adoption, Railpen’s report introduces how AI systems can be classified, outlines the key risks they pose to companies across different sectors and roles in the AI value chain, and presents an AI Governance Framework (AIGF) that translates responsible AI principles into actionable practices.
The report provides insights on:
• Key short-term AI risks for businesses and how these risks have the potential to financially impact businesses by increasing the legal, regulatory, operational and reputational risks that they face. These include
o social risks (e.g., through AI’s potential impacts on employment and labour, intellectual property, misinformation, privacy and digital rights, and bias and discrimination),
o governance risks (e.g., through AI’s potential impacts on accountability, transparency and decision-making),
o environmental risks (e.g., through AI’s potential impacts on energy consumption and resource use).
• Why investors need to take a governance-based approach to AI management, due to high uncertainty around its medium and long-term implications. For example, in terms of how AI will affect market structures, sectors, and individual companies. The report proposes a high-level AIGF to assess companies’ approaches to AI, with expectations for
o governance (e.g. senior or board-level oversight, management and policies),
o strategy (assessing AI’s relevance to the business strategy),
o risk management (e.g., identification and monitoring of AI risks, management of AI risks, stakeholder engagement)
o performance rating (e.g., AI incident reporting, annual reporting).
· The ways investors have begun to address AI risks through publishing their governance expectations, engaging with companies, voting, and participating in collaborative initiatives.
· Recommendations and guidance for investors on company engagement and policy advocacy towards effectively assessing and managing AI risks.
The report also highlights how Railpen is seeking to ensure members of the railways pension schemes benefit from the tailwinds offered by AI adoption, across both its growth-oriented, bottom-up fundamental equities portfolio and through its real assets investments.
Chronos’ view:
Although AI reflects traditional ESG risks in many respects, the very scale, speed and complexity of developments in AI that makes it such a powerful and positive disruptor, makes any associated risks more difficult to predict and manage for all stakeholders, particularly regulators. This means investors must ensure that companies have the systems and processes needed to ensure AI is harnessed in a safe and responsible manner.
As emphasised by the report, the potential long- and medium-term risks of AI are highly unpredictable and virtually unknown so investors must focus on the governance of AI rather than specific risks in order to manage long-term risk – and ensure company boards can effectively harness AI opportunities too. They should ensure the companies they invest in have adequate risk management in place by taking a governance-based approach to managing AI risks, focusing on companies’ long-term governance structures while remaining abreast of the evolving short-term risks.
To access the report, please see here.
