HealthQuill Climate AI’s energy demands linked to asthma risks, premature deaths
Climate Health Medical Research

AI’s energy demands linked to asthma risks, premature deaths

A ransomware group has breached the information security network of India's Sun Pharmaceuticals unit, and there is a possibility that "file systems, business and personal data" might have been compromised.

AI's energy demands linked to rising health risks: Alarming predictions for asthma and death rates by 2030

Bharti Jayshankar

December 11, 2024: The rapid expansion of data centres driven by artificial intelligence (AI) is raising serious public health concerns, with predictions indicating a significant rise in asthma cases and premature deaths due to air pollution.

Researchers from the University of California, Riverside, and California Institute of Technology have revealed that by 2030, data centres could contribute to an additional 600,000 asthma cases and 1,300 premature deaths annually in the United States. This would account for more than one-third of asthma-related deaths in the country.

Shaolei Ren, a lead researcher, emphasizes that these impacts are not confined to local areas; airborne pollutants can travel long distances, affecting communities nationwide. The researchers say that the estimated public health cost due to fueling AI could be over $20 billion annually by 2030.

A Goldman Sachs research estimates that data centre power demand will grow 160% by 2030. At present, data centres consume 1-2% of overall power across the world, which will double in a decade. The carbon dioxide emissions of data centres may more than double between 2022 and 2030.

The last decade has seen the demand for data consumption almost triple, but there was o significant energy consumption spike as the process brought in some energy efficiency too. However, the recent jump in AI and data-driven consumption has led to huge energy consumption with no energy offset.

A single ChatGPT query requires 2.9 watt-hours of electricity, compared with 0.3 watt-hours for a Google search, according to the International Energy Agency.

Sources of pollution

The electricity required for AI operations often comes from fossil fuels, leading to the emission of harmful pollutants such as fine particulate matter and nitrogen oxides. For instance, the energy consumption for training a large AI model could generate air pollution equivalent to driving a car over 10,000 round trips between Los Angeles and New York City.

In Virginia’s Data Center Alley, backup diesel generators are particularly problematic. Even at just 10% of their permitted emissions levels, these generators could already be causing 14,000 asthma symptoms annually and imposing public health costs ranging from $220 million to $300 million per year. If emissions reach maximum permitted levels, costs could skyrocket to between $2 billion and $3 billion annually.

Broader implications

The implications extend beyond immediate neighborhoods. Pollution from data centres can affect low-income communities disproportionately due to their proximity to power plants and backup generators. Ren notes that while data centres contribute economically through local property taxes, they do not compensate communities affected by their pollution.

Future of data centres and energy consumption

As AI continues to evolve, so does the energy demand from data centres. Projections indicate that by 2030, data centres could consume between 11% to 12% of total electricity in the U.S., up from just 3% to 4% last year

This surge is expected to double the carbon emissions associated with data centres during this period.

The findings highlight an urgent need for tech companies to adopt cleaner energy sources and improve reporting on air pollution linked to their operations. While some companies are investing in renewable energy projects and nuclear power technologies, many still rely heavily on fossil fuels. The researchers call for standards that require accountability from tech companies regarding their emissions and suggest compensating affected communities.

As the demand for AI services grows, understanding and mitigating the public health impacts associated with data centre emissions will be crucial for ensuring a healthier future for all.

Exit mobile version