California and the EU Move to Rein In the Environmental Cost of AI

A global crackdown on the water and energy use of the artificial intelligence sector is picking up speed. This week government officials in California and the European Union both announced new policies to help track and limit the runaway resource use of hyperscalers as the threat of AI’s environmental impact looms ever larger.By the end of this decade, artificial intelligence will consume as much water as 1.3 billion people in sub-Saharan Africa, and triple the energy consumption of Pakistan, Bangladesh, and Nigeria put together – a population of 650 million people. Plus, its land footprint will be massive, potentially exceeding 14,500 square kilometres. That’s according to a recent study from United Nations University, which found that artificial intelligence is increasingly sucking up water, energy, and land resources that are gravely needed by the growing and developing global population.“This report is not a case against artificial intelligence, a technological transformation that is improving the lives of billions of people around the world,” said Professor Kaveh Madani, who led the investigation team for the United Nations University study. “It is a call for using it responsibly and addressing its unintended impacts proactively to make it sustainable and equitable. We have a narrow window to ensure that the backbone of the technological revolution of our era develops within planetary limits, and that the communities who provide the critical minerals for advancing AI and the ones that host its infrastructure and e-waste are also among those who benefit from it.”Policymakers are increasingly answering this call, attempting to track and rein in the sector’s out-of-control resource use. On Monday, the European Commission, which functions as the executive branch of the European Union, introduced new rules requiring data centers to disclose their energy and water usage. It also requires data centers to report the relationship between their own water usage and local water stress. This transparency would come at a critical juncture for Europe, as the bloc aims to triple its data centre capacity over the next seven years in order to remain technologically independent and competitive with the United States’ artificial intelligence sector.While these provisions do not set a limit on the amount of energy and water data centers can use, they may lay the groundwork for future legislation that does just that. Part of the problem with regulating the global artificial intelligence sector is that it operates under extreme opacity. It’s difficult to quantify exactly how much energy and water the sector is already using, much less write models for how to responsibly scale that use. As such, transparency and labelling schemes are an important first step toward creating good policy.Meanwhile, in California, Governor Gavin Newsom has signed a flurry of new bills – seven in total – imposing similar provisions on the state’s data centers. The package of laws requires proposed data center plans to estimate their water usage and include data on energy efficiency and drought contingencies. It also sets requirements for energy, water, and fuel consumption, which data center proposals must meet in order to move forward in the official approval process. In addition, the new provisions require data centers to fund much-needed grid upgrades in the localities where they will be constructed.“While the Trump administration moves toward deregulation, communities are left to deal with the consequences — higher electricity demand, grid constraints, water use, and pollution,” Newsom said in a press release accompanying the announcement of the new laws. “With these laws, we are ensuring that Californians remain in the driver’s seat — and that those profiting from data centers aren’t doing so at our expense.”While Newsom’s comments speak to the political and regulatory climate in the United States, this argument reflects a broader imbalance that artificial intelligence is spreading on a global scale. While the world is benefiting in many ways from the proliferation of artificial intelligence (but not without enormous and terrifying trade-offs), the costs of its expansion are unduly carried by certain regions.The United Nations warns of a growing digital and environmental divide, where the United States and China control over 90 percent of artificial intelligence computing capacity, while many other nations are set to suffer from the environmental costs of that computing without benefiting from the economic development that it can usher in.By Haley Zaremba for Oilprice.com More Top Reads From Oilprice.comOil Extends Rally Despite Higher Hormuz Volume ReportsChina's LNG Imports Set for Second Straight Monthly DropWorld's Tallest Dam Project Gets New Life With Fresh World Bank Funding