Key Takeaways

The discussion highlighted the need to move beyond the question of what artificial intelligence can do towards a more fundamental question: what should AI be used for, and under what conditions does it create genuine value?

A speaker underscored that organisations should first identify their needs and determine which tasks would benefit from AI. This is especially important as, according to McKinsey, 88% of companies use AI, but only 6% of them see value in it. As a result, companies should have a deeper understanding of the problems they face and the tools needed to solve them, as different tasks require different AI solutions.

Another speaker reminded participants of AI’s tremendous resource consumption. As AI requires significant infrastructure to function, the pressure on infrastructure investment was raised during the discussion, especially given that artificial intelligence is already very expensive.

AI was also discussed as a systemic risk, as despite rising capital investment and job growth in infrastructure, it still creates uncertainty around resource consumption, employment and geopolitical risks.

Gaps & Opportunities

A significant challenge remains: sustainability comes second when considering investments in or designs for AI. Organisations still prioritise performance, growth and competitive advantage without fully accounting for environmental externalities, including energy and water consumption.

Another gap is that many companies are still learning how to integrate AI effectively into their operations. Without clear objectives and appropriate governance, organisations risk investing in expensive systems that deliver limited practical value. Furthermore, frameworks for applying “sustainable AI” remain limited, as sustainability is still a broadly defined term. What is meant by sustainable materials may also vary from one place to another, given differing environmental, economic and political circumstances.

The discussion also raised questions about accountability: who is responsible for the environmental and social consequences of AI, and who should establish the standards for its sustainable use, governance or education?

However, opportunities remain. More sustainable AI technologies already exist, including advanced chips, better cooling systems and grid optimisation. These could help balance the benefits AI brings with its resource consumption.

Investors also have an opportunity to influence the development of AI. Through shareholder engagement and capital allocation, they can encourage companies to improve transparency, manage resource consumption and demonstrate that AI investments create long-term value.

Recommendations

• Organisations should adopt a purpose-driven approach to AI. Before investing, they should define the problem and evaluate which tasks require AI.
• Sustainability should be considered to a greater extent when investing in AI. Companies should assess energy and water use, emissions, material requirements and social consequences alongside financial returns and productivity gains. Transparent reporting would make these impacts easier to compare.
• Stronger collaboration is needed between governments, businesses, researchers and investors. Common standards for responsible and resource-efficient AI could clarify accountability and help address risks that individual organisations cannot manage alone.
• Investors should treat AI as both an opportunity and a systemic risk. Engagement with companies should encourage responsible innovation, appropriate governance and investment aligned with long-term needs.

The future of sustainable finance will depend on its ability to direct capital towards technologies that create genuine economic and social value without exceeding planetary boundaries.