• The cooling systems and generators of AI data centers are causing illnesses
• Noise emitted from data centers is below the threshold of human hearing
• The infrasound can be 'felt', and is causing dizziness, nausea, anxiety, and more

People living nearby to AI data centers in the US are increasingly reporting illnesses caused by a near-imperceptible hum.

The infrasound, which in some cases can be ‘felt’ rather than heard, is causing people living in the vicinity of several data centers to fall sick with symptoms such as headaches, insomnia, nausea, and anxiety.

Those living near certain data centers have reported noise levels approaching 100dB, 24 hours a day, seven days a week.

Undetectable hum

Numerous grassroots groups and local communities have launched bids to stop or halt the construction of data centers on the grounds of electricity consumption, pollution, and general opposition to AI.

Now, noise pollution concerns could soon join the list. According to the Environmental and Energy Study Institute (EESI), data centers emit sounds at a range of frequencies, both high and low. Lower frequencies, especially those below the threshold of human hearing, are particularly difficult to detect without reliable equipment, making noise auditing for local communities especially difficult.

Residents of Brittany Heights in Chandler, Arizona, were forced to try and block out the constant hum of a data center built in 2014 using noise-cancelling headphones and earplugs with little effect. The noise pollution was such a big issue for local residents that it led to the proposal for a new data center being blocked.

As AI data centers rely on huge ranks of GPUs to run, they generate an enormous amount of heat. The GPUs require cooling to run at peak efficiency, meaning that huge amounts of energy are used to cool the air. This can amount to up to 40% of a data center's total electricity consumption.

Many data centers also make use of backup generators to keep their electricity supply at 100%. In many cases, these are diesel generators that are switched on during peak times when electricity suppliers need to provide power to other residential and industrial areas. As a result, banks of generators releasing as much as 105dB each are switched on to keep the power flowing.

It is even worse for off-grid data centers. Rather than building the expensive infrastructure to connect a data center to the existing infrastructure near population centers, some data centers are built in rural areas and powered by natural-gas-powered turbines. These turbines generate electricity in a similar way to how jet engines generate thrust. The sound of these turbines can be heard from many miles away.

A study published in the US National Library of Medicine tested the effects of infrasound above 100dB on human tissue. The experiment concluded that infrasound can affect cardiac function in as little as one hour after exposure.

Infrasound has been also referenced as a potential cause of the so-called ‘Havana Syndrome’ experienced by US and Canadian diplomats and their families working at certain locations abroad, who experienced similar symptoms to those living near data centers.

Via Tom’s Hardware

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.

Benedict is a Senior Security Writer at TechRadar Pro, where he has specialized in covering the intersection of geopolitics, cyber-warfare, and business security.

Benedict provides detailed analysis on state-sponsored threat actors, APT groups, and the protection of critical national infrastructure, with his reporting bridging the gap between technical threat intelligence and B2B security strategy.

Benedict holds an MA (Distinction) in Security, Intelligence, and Diplomacy from the University of Buckingham Centre for Security and Intelligence Studies (BUCSIS), with his specialization providing him with a robust academic framework for deconstructing complex international conflicts and intelligence operations, and the ability to translate intricate security data into actionable insights.