Regulating the Energy Wave in the Age of AI and Data

Growing focus on AI regulation

The digitisation of society and industry continues to accelerate at an unprecedented rate. AI is opening the door to advances right across society, from healthcare to agriculture. Financial institutions are integrating and exploring AI models across the board, in areas such as cybersecurity, risk management and AML processes. As a result, we’ve seen increasing regulatory focus on this area. Of particular significance is the recent EU AI Act, which places an emphasis on protecting fundamental rights.

Growing energy demands

This growing quantity of digital data requires an expansion of data centres to store and process it. The International Energy Agency estimates that the energy consumption driven by data centres, cryptocurrencies and AI consumed about 2% of the total global electricity demand in 2022 and could double towards 2026. Climate change has been at the forefront of regulatory priorities for several years now; however, the increasing energy demands of AI poses some fundamental challenges to this progress. Regulators are grappling with the pursuit of climate targets alongside the need to foster innovation and harness the sustainability and climate change solutions and other societal benefits that AI has to offer.

Regulatory measures

Often setting the pace in environmental regulation, the EU has regulated in this specific area, with concrete sustainability reporting requirements for companies in relation to data centres under the Corporate Sustainability Reporting Directive (CSRD) and the recast Energy Efficiency Directive (EED). The Commission has approved a delegated regulation specifically establishing a reporting scheme to rate the sustainability of EU data centres. Forming part of the ‘Fit for 55’ legislative package and the recast Energy Efficiency Directive, it requires data centre operators to report the key performance indicators to the European database by 15 September 2024 and then by 15 May in subsequent years.

Sustainability reporting is proving to a polarizing issue in the USA, with proposed federal measures facing significant resistance. Against this background, the Artificial Intelligence Environmental Impacts Act of 2024 was introduced as a bill by Senator Markey in February of this year. Among the measures proposed under the Act is a requirement for the National Institute of Standards and Technology (NIST) to develop standards to measure and report AI’s environmental impacts as well as the creation of a voluntary framework for AI developers to report environmental impacts. At state level, an example of reporting requirements has been introduced in California, which often leads the way in climate related measures. Two acts have been passed: the Climate Corporate Data Accountability Act (SB 253) and the Greenhouse Gases: Climate-Related Financial Risk Act (SB 261), which will require many large companies to make climate-related disclosures starting in 2026. 

In 2019, Singapore imposed a moratorium on the construction of new data centres, with concerns around the energy requirements and carbon footprint driving this decision. This was lifted in 2022 with measures in place to ensure that data centres are energy efficient. In June 2023, Singapore launched one of the world’s first standards for optimising energy efficiency for data centres (DCs) in tropical climate countries. The Infocomm Media Development Authority’s standard forms part of the Digital Connectivity Blueprint which puts sustainability at the heart of Singapore’s digital infrastructure.

Shifting regulatory priorities

We are only beginning to understand AI’s full potential in relation to climate change management and solutions. There are numerous examples of its use already - in a January 2024 paper the World Economic Forum identified nine ways in which AI is helping tackle climate change, which cover monitoring tools that predict weather, track icebergs and identify pollution. There are numerous ways that AI tools can and are being used by financial firms for data collection and accurate emissions reporting. However, difficult decisions will need to be made as we seek to balance the need to regulate AI energy use with its potential benefits for society and the environment as a whole. Regulation needs to drive sustainable practices, ensure transparency and encourage innovation. It will need to come up with a method of not only regulating the risk of harm, but also evaluating the potential for benefit.