California Aims to Address Air Pollution Associated with Artificial Intelligence

December 17, 2024 TruthUSA

A recent study indicates that emissions from artificial intelligence (AI) could surpass those of all 35 million vehicles registered in California by 2030.

On December 16, California Governor Gavin Newsom remarked that the state is encountering distinct challenges pertaining to energy and natural resource usage as a leading hub for artificial intelligence.

The governor pointed out that 32 out of the top 50 AI firms globally, including OpenAI, Anthropic, and Databricks, are based in California.

“This is a world that’s radically changing,” Newsom stated in an interview with The Epoch Times. “California is at the forefront of this evolution.”

According to a newly published study, the rise in energy consumption to meet AI demand is escalating air pollution, and existing risk models do not adequately account for these costs.

“We aim to remain competitive in this area, but we are also aware of the energy implications,” Newsom added.

He proposed that emerging technology such as small modular nuclear reactors—which generate carbon-free nuclear energy using more compact facilities that require lower investments than traditional nuclear power plants—could assist the state in reaching its objectives.

“There is currently a race regarding energy, particularly in relation to [AI computing] and its components,” Newsom remarked. “This is an evolving challenge for California as we continue to excel and lead as the primary beneficiary of this industry while also pushing forward our low-carbon growth initiatives.”

Researchers from the University of California–Riverside and the California Institute of Technology have discovered that AI is adversely impacting public health and is “significantly” exacerbating air quality issues by increasing the number of fine particulates released into the atmosphere through nitrogen and sulfur dioxide emissions.

“These are harmful pollutants that can penetrate deep into people’s lungs, yet we do not typically associate them with AI computing. When we think of air pollution, cars are usually what comes to mind,” explained Shaolei Ren, an associate professor of electrical and computer engineering at UC Riverside, in a statement to The Epoch Times.

The study indicates that training AI models generates pollution equivalent to driving a round-trip from Los Angeles to New York City over 10,000 times.

In 2023, the public health impacts attributed to AI were comparable to about one-third of the 35 million vehicles on California roads; by 2030, emissions from AI could exceed those from cars, according to the findings.

The research also suggests that the industry forecasts nearly tripling electricity consumption for data centers powering AI initiatives over the next seven years, potentially representing nearly 12 percent of total energy usage nationwide by 2030.

The escalating need for data centers is exerting pressure on power grids, negatively affecting the environment through increased water usage and the proliferation of particulates in the atmosphere, as concluded by the study alongside another study led by Ren and colleagues in 2023.

The authors recommend setting criteria to report air pollution impacts and balancing workloads to optimize solar energy production and other alternative energy sources that could alleviate health issues.

“We aren’t advocating against AI development, but similar to any decision-making process, we need to understand both benefits and costs. Currently, the emphasis seems to be on the benefits while the immediate health impacts are overlooked,” said Ren. “It is crucial that we consider the net costs so that we can make more informed choices.”

The implications of the AI sector on public health represent what researchers label a “hidden toll.”

The study estimates that pollution linked to AI could result in around 600,000 asthma cases and 1,300 premature deaths by 2030, with total annual public health costs possibly exceeding $20 billion.

Exposure to air pollution has strong associations with asthma, cardiovascular diseases, heart attacks, and premature death, among other health concerns.

The World Health Organization indicated that approximately 4.2 million deaths globally were attributed to air pollution in 2019, marking it as the second leading cause of noncommunicable diseases.

According to the American Lung Association’s State of the Air 2024 report, 40 percent of the U.S. population resides in regions with unhealthy particulate levels in the air.

Pollution affects communities nationwide, as emissions can travel hundreds of miles across state lines, according to the findings of the AI study.

Data centers require constant operation, resulting in significant power consumption. Some of this energy is sourced from fossil fuel-powered plants, and diesel generators—which discharge up to 600 times more particulates than natural gas plants—are often employed for backup, further contributing to air pollution, the researchers noted.

The authors advocated for a thorough cost-benefit analysis that encompasses both short- and long-term health effects.

Potential advantages of AI include enhanced efficiency, increased productivity, and lower costs, as identified in various university studies.

Recognized as the fastest-growing tech sector and a pivotal innovation across multiple industries, AI is poised to play a “significant” role in the lives of countless individuals, remarked Adam Wierman, Caltech’s Carl F. Braun professor of computing and mathematical sciences and director of information science and technology.

“At the same time, we must ensure that we address the negative repercussions that accompany it, and that they are duly acknowledged, assessed, minimized, and equitably shared,” he asserted in a statement.

Source link