PIMFA
AI and the Environment: A Double-Edged Sword for the Investment Industry
Artificial Intelligence (AI) is set to have major impact on the investment landscape. Yet, its adoption remains a topic of debate. On one side, critics voice concerns about how AI might disrupt the workplace of wealth managers and advisers. On the other, proponents highlight its potential to boost productivity, reduce human error, and enhance cost-efficiency through automation and speed, and its potential to enhance the client experience. However, the promise of AI stands in a delicate balance with the growing emphasis on sustainability in investment practices.
A recent survey revealed that 62% of wealth management firms believe AI will significantly reshape their operations. This shift aligns with investor expectations for digital experiences on par with leading technology firms. Yet, the infrastructure powering AI, such as data centres, poses a substantial environmental challenge, generating vast amounts of e-waste and consuming significant energy.
The Promise and Paradox of AI in Sustainable Investing
AI could be immensely useful for sustainable investing. Its ability to verify and analyse data rapidly is a game changer, offering insights by processing vast quantities of structured and unstructured data into reliable, actionable information. This efficiency could support ESG (Environmental, Social, and Governance) investing by helping firms identify sustainable opportunities with precision.
However, the environmental cost of AI should be considered. Training advanced AI models often results in a significant carbon footprint. For instance, training a large language model (LLM) can generate more emissions than 125 round-trip flights between Beijing and New York[1]. This stark contrast between AI’s potential benefits and its environmental toll warrants closer scrutiny.
The Hidden Environmental Costs of AI
When firms are considering the adoption of AI solutions, they will need to start taking account of their impact on the environment. On the face of it, AI looks like a clean, environmentally friendly resource, but there is an environmental cost which needs to be considered.
Beyond carbon emissions, AI’s environmental impact extends to water consumption, a resource essential to both AI infrastructure and operations. Some of the aspects to consider are:
- Chip Manufacturing: Production of microchips, the backbone of AI systems, requires approximately 2,200 gallons of ultra-pure water. Semiconductor factories, which produce these chips, consume as much water annually as 7.5 million people living in Hong Kong, according to the World Economic Forum[2].
- Data Centre Operations: The running and cooling of AI systems in data centres demands substantial water resources. For example, OpenAI’s GPT-3 model requires millions of litres of fresh water for training and operation. A single user query can consume between 10 to 50 millilitres of water, depending on the hosting location[3].
- Corporate Impact: Major tech companies have reported surging water usage. In 2022, Google’s on-site water consumption rose by 20%, while Microsoft’s jumped by 34%, partly due to the demands of AI systems, according to OECD.AI.
- Energy consumption: AI requires a lot of energy, which is often generated by burning fossil fuels, a major contributor to global warming. The International Energy Agency estimates that by 2026, AI, cryptocurrency, and data centres could use 4% of the world’s annual energy.
- Electronic waste: Data centres that house AI servers produce electronic waste that contains hazardous chemicals like lead, mercury, and cadmium.
Navigating the Environmental Dilemma in Wealth Management
For firms committed to sustainable investing, addressing their own environmental impact is crucial. As people are looking to align their investments with their personal values, clients will expect both financial returns and a commitment to sustainability from wealth managers. Firms must consider their technology stack and operations alongside their investment portfolios.
Efforts are underway to mitigate the environmental impact of AI. Companies are exploring energy-efficient AI models, adopting renewable energy for data centres, and improving hardware recycling processes. Wealth management firms should actively evaluate these alternatives as part of their sustainability strategies.
AI presents both challenges and opportunities for the investment industry. By harnessing its potential while addressing its environmental costs, firms could position themselves as leaders in responsible innovation. The road ahead demands a thoughtful balance between leveraging AI’s transformative capabilities and honouring commitments to a sustainable future.
The Regulators Actions
The FCA is collecting more information about AI to look at the risks and opportunities AI presents to UK consumers and markets and help inform the regulatory approach in a practical way. For that all stakeholders are invited to provide their view on the future of AI and what is possible and sensible within financial services. More information on the FCA initiatives can be found here.
Maria Fritzsche –Â Senior Policy Adviser, PIMFA
[1] https://eprints.gla.ac.uk/324195/1/324195.pdf
[2] https://www.weforum.org/stories/2024/07/the-water-challenge-for-semiconductor-manufacturing-and-big-tech-what-needs-to-be-done/#:~:text=This%20burgeoning%20demand%20for%20chips,to%20an%20S%26P%20Global%20report
