AI Governance and ESG: Why Responsible AI Is Becoming a Sustainability Issue
Governance has always been the 'G' in 'ESG,' the part that covers how an organisation makes decisions, who is accountable for what, and whether structures exist to ensure the business acts responsibly over the long term.
Until recently, most governance discussions in ESG typically focused on board composition, executive remuneration, anti-corruption practices and supply chain conduct.
However, Artificial Intelligence is adding a new dimension to that conversation. Organisations deploying AI tools are increasingly being asked, by investors, by regulators and by their own leadership, how they are governing that use. And the connection between AI governance and environmental and social sustainability is more direct than it might first appear.
The intersection of AI and ESG operates on multiple levels. There is the environmental impact, the energy and water consumed by the data centres that power AI tools, and the carbon emissions associated with AI workloads.
There is the social dimension, how AI affects employment, how algorithmic decision-making intersects with fairness and inclusion, and what it means for organisations' relationships with their employees and communities.
And there is the governance dimension, how organisations decide which AI tools to use, what oversight exists for AI-driven decisions, and how they ensure AI use is consistent with their stated values.
ESG frameworks are still catching up with AI. But the direction of travel is clear: organisations that have given serious thought to how they govern and account for their AI use will be in a better position than those that haven't, as these expectations develop and crystallise into formal requirements.
As covered in our companion article on AI's environmental footprint, the energy and water consumed by AI workloads is not trivial. For organisations with sustainability commitments, this creates a practical challenge: AI use generates emissions that belong within the organisation's Scope 3 carbon footprint, under the purchased goods and services category, yet most organisations are not tracking or reporting it.
The corporate responsibility here is analogous to how organisations handle other Scope 3 emissions. Just as procurement teams increasingly consider the environmental credentials of their suppliers, sustainability teams are beginning to consider the environmental credentials of their digital service providers, including AI tool providers.
Questions worth asking include: Where are the data centres that power the AI tools we use? Are they powered by renewable energy? How do the providers account for and report their energy use? Is this information available in a form that can be included in our own carbon reporting?
These questions don't yet have universally available answers. But organisations that start asking them are developing a discipline that will matter as Scope 3 reporting becomes more rigorous.
The social and governance dimensions of AI are particularly relevant in the context of ESG reporting frameworks, which are increasingly asking about how organisations use data and technology in ways that affect people.
Algorithmic decision-making, where AI tools inform or automate decisions about employees, customers or suppliers, raises questions about fairness, transparency and accountability. If an AI system influences how performance reviews are conducted, how credit is assessed, how job applications are screened or how pricing is set, there are legitimate questions about whether that system has been designed and tested for fairness, whether its outputs are explained to the people affected, and whether there is human oversight of the decisions it influences.
These are governance questions; they are about accountability structures and decision-making processes, not just technology choices. They sit squarely within the G of ESG, and they are beginning to appear in ESG disclosure frameworks and investor questions.
For most UK & Irish organisations, AI governance is not yet a formal programme, and it doesn't need to be to be worth thinking about. The starting point is a straightforward internal review of how AI tools are being used across the organisation, what decisions they are informing or automating, and whether appropriate oversight exists.
Practical governance measures that organisations are beginning to put in place include:
None of these steps requires a major investment or a dedicated AI team. They require the same kind of systematic thinking that good governance has always required, identifying risks, establishing accountability and putting in place the processes to manage both.
AI regulation is developing rapidly in Europe. The EU AI Act, which applies from August 2024, introduces risk-based requirements for AI systems, with the most stringent requirements applying to high-risk applications. Irish businesses deploying AI tools in areas like recruitment, credit assessment, critical infrastructure or public services will need to understand how the AI Act applies to their specific use cases.
Beyond the AI Act, the connection between AI governance and sustainability reporting is likely to develop through CSRD and related frameworks, as the social and environmental impacts of AI become better understood and more consistently defined. Organisations that have already put thoughtful AI governance structures in place will find it easier to demonstrate compliance as these requirements develop.
The purpose of AI governance in an ESG context is not to restrict AI use or to treat it with suspicion. AI offers genuine benefits, for efficiency, for insight, and increasingly for sustainability work itself. The purpose is to ensure that AI use is intentional, accountable and consistent with the organisation's stated values.
An organisation that has committed to reducing its carbon footprint should be asking whether its AI tool usage is contributing to that goal or working against it. An organisation committed to fair treatment of employees should be asking whether AI tools in its HR processes are operating consistently with that commitment. These are not uncomfortable questions, they are exactly the kinds of questions that good ESG governance asks about every significant aspect of an organisation's operations.
What is AI governance in the context of ESG?
AI governance in ESG refers to the structures, policies and processes an organisation puts in place to ensure its use of artificial intelligence is responsible, transparent and accountable. This includes oversight of how AI tools are used in decision-making, how the environmental impact of AI use is measured and reported, and how AI-related risks, including bias, data privacy and regulatory compliance, are managed.
Why is responsible AI becoming a sustainability issue?
Because AI use has real environmental, social and governance implications. The energy and water consumed by AI data centres generates carbon emissions that belong in an organisation's Scope 3 footprint. AI tools that inform decisions about people raise fairness and transparency questions relevant to the social dimension of ESG. And the governance structures around AI use are increasingly expected to be part of how organisations demonstrate responsible corporate behaviour.
How do organisations include AI governance in their ESG policies?
Practical starting points include maintaining an inventory of AI tools in use, establishing accountability for AI-related decisions, including AI providers in sustainability supply chain assessments, and incorporating AI governance into board-level ESG reporting. Many organisations are beginning to address AI governance through their existing data protection, procurement and risk management frameworks, rather than creating separate AI-specific structures.
What does the EU AI Act mean for Irish businesses?
The EU AI Act, applying from August 2024, introduces risk-based requirements for AI systems. High-risk applications, including AI used in recruitment, credit decisions, critical infrastructure and public services, face the most stringent requirements around transparency, human oversight and documentation. Irish businesses using AI tools in these areas should understand how the Act applies to their specific use cases.
Should AI use be included in carbon footprint reporting?
Increasingly, yes. AI-related emissions fall within Scope 3 under purchased goods and services. As Scope 3 reporting becomes more rigorous under CSRD and related frameworks, the emissions associated with AI tool use are likely to come under greater scrutiny. Organisations that are already tracking their digital and AI-related emissions will be better prepared and will have a more complete and credible picture of their total footprint.
If your organisation is beginning to think about how AI use fits into your ESG strategy and governance framework, AD Sustainability can help you develop a practical, proportionate approach that addresses both the environmental and governance dimensions. Get in touch today.