AI as a force for good in combatting climate change

Scientists are developing AI solutions to support efforts to meet climate targets and help slow down and reverse environmental impacts. Whilst there is unlikely to be a ‘silver bullet’ solution to climate change, the combined effect of initiatives that use AI to enhance insights about the cause and effects of climate change and to optimise energy consumption may be very significant. In fact, AI may play a key role in addressing sustainability challenges. Some examples of ongoing initiatives include:

• iceberg mapping capabilities developed by researchers at the University of Leeds which process satellite images to track iceberg movement and changes;

• use of modelling by the United Nations AI Advisory Body to predict weather patterns that can help communities and authorities to prepare for adverse weather events;

• use of data by logistics firms to optimise freight transportation routes;

• WWF’s Forest Foresight, which uses modelling to predict forest loss and help forest conservation efforts; and
• use of AI to optimise the energy grid, reduce energy waste and enhance the efficiency of heating and cooling facilities.

The environmental impact of AI technology

At the same time, AI is extraordinarily energy intensive. A generative AI system is estimated to use around 33 times more energy than simple task-specific software. The consumption of electricity by data centres has dramatically increased in recent years. A large data centre can consume as much power as 80,000 households, and it is predicted that by 2025, data centres could account for up to 3.2% of total carbon emissions. While other factors have contributed to this (cryptocurrency, for example, involves significant energy consumption) the vast volumes of data processed by generative AI systems have caused a surge in energy consumption.

It's not just power consumption that is an issue. Data centres consume vast quantities of water, primarily for cooling infrastructure and equipment. In addition, many AI technologies use hardware that is manufactured using finite minerals and resources, such as lithium, cobalt and tantalum. The extraction and use of natural resources impacts the environment and local communities.

This kind of resource demand, while critical to the functionality of AI and other emerging technologies, conflicts with global efforts to address climate change.

As the environmental impact of AI becomes more evident, regulators and policymakers worldwide are taking notice and taking action. Governments in countries such as Singapore, Germany and the Netherlands are imposing sustainability standards on new data centre constructions. In the UK, amendments to the Minimum Energy Efficiency Standards (MEES) and UK Building Regulations in England impose conditions on the grant of commercial leases (including data centre leases) and require new buildings (including data centres) to meet requirements on CO2 emissions.

The impact on insurance?

For the insurance industry, the impact of climate change is profound, and AI may offer solutions to some of the most pressing challenges for insurers. The increasing frequency and impact of natural disasters, combined with pressures to meet changing regulatory requirements, present a threat to traditional insurance business models as climate-related risks become increasingly difficult and/or expensive to underwrite.

AI solutions for assessing and pricing risk will be critical to insurers offering products in areas where environmental events present significant risk. For example, AI can be leveraged to better assess the risk (and even predict) the occurrence of floods or forest fires.  Claims automation capabilities powered by AI can also assist insurers to manage large volumes of claims arising from adverse weather events. At a portfolio level, insurers can also use AI to more effectively stress-test their exposure to climate risk and adapt their products and their business models accordingly.

Insurers also face pressure to meet their own ESG commitments and regulatory requirements. The EU Corporate Sustainability Reporting Directive (CSRD) was introduced on 5 January 2023, imposing comprehensive sustainability reporting standards on insurers. Under pressure from investors, customers, and regulators, insurance companies will need to assess the environmental impacts of the technologies they adopt and may need to consider carbon offsetting measures to mitigate the environmental impact of their data centres and data processing capabilities.

Concluding comments

AI offers tremendous potential to support global climate and sustainability efforts. For the insurance industry, the prominence of climate risks has been elevated by the increasing frequency of climate events, together with increasing pressure to meet ESG commitments and regulatory requirements. AI will play a key role in enabling insurers to adapt their business models and their products considering increasingly volatile climate risks.

However, the impact of AI technologies on the environment cannot be ignored. There is increasing pressure on governments and businesses to implement measures to mitigate and offset the energy and resource consumption of AI capabilities. Insurance companies, like all other businesses, will need to find a way to balance  the benefits of AI with their sustainability objectives in order to ensure that the adoption of this powerful technology does not come at the expense of climate commitments.

Authored by Lydia Savill, Louise Crawford and Rita Hunter.