Experts Warn of Hidden Dangers from AI Data Centers
As artificial intelligence (AI) technology rapidly evolves, health and environmental advocates are raising alarms about its potentially hazardous side effects. They warn that data centers driving the AI boom may be releasing harmful substances known as “forever chemicals” into surrounding communities. This issue has received insufficient attention, according to experts.
The rise of AI has been widely celebrated for its potential to transform various industries, enhancing energy efficiency, monitoring environmental changes, and improving food supply management. Yet, these advancements come at a cost. Data centers, which power AI operations, consume significant energy and water, leading to concerns about their environmental impact.
Recent reports by The Guardian highlight that major tech companies, including Google, Microsoft, and Amazon, are utilizing per- and polyfluoroalkyl substances (PFAS), also referred to as f-gas, for cooling their data centers. These chemicals could pose serious environmental and health risks as they may leak into the atmosphere or contaminate land when equipment is disposed of. The growing demand for these technologies has led to increased production of f-gases, with companies like Chemours ramping up output in their West Virginia and North Carolina facilities to meet industry needs.
Jonathan Kalmuss-Katz, an attorney with Earthjustice, emphasizes that this matter has been “dangerously understudied,” raising concerns about the lack of regulatory oversight. As data center operations expand, so too could the exposure to these chemicals, which have been linked to severe health issues, including reproductive problems and cancer.
PFAS are notorious for their persistence in the environment, earning them the “forever chemicals” label. They do not degrade easily, leading to potential generational health impacts. Alarmingly, companies are not required to disclose their f-gas usage, and regular testing at data center sites remains absent. Industry claims suggest that escaping f-gas does not pose a threat because it transforms into a compound called Tfa, which is not classified as a PFAS in the United States, though it is recognized as such in other countries.
The Environmental Protection Agency (EPA) has recently indicated it would expedite chemical reviews to facilitate data center operations, a move that has raised concerns among advocates. EPA Administrator Lee Zeldin stated that the agency aims to “get out of the way and help speed up progress,” which many interpret as a shift toward less stringent oversight.
Maria Doa of the Environmental Defense Fund criticized this approach, arguing that the EPA’s strategy could undermine public safety. She noted that insufficient regulations on data centers using potentially harmful chemicals might also endanger workers in those facilities and the plants producing their equipment.
In light of these challenges, some states, such as Minnesota, are advocating for legislation that requires tech companies to disclose their chemical usage. As the conversation surrounding AI regulations gains momentum ahead of the U.S. mid-term elections, constituents are likely to push candidates to address public health risks and environmental concerns associated with the industry.
Until stricter regulations are implemented, residents may find themselves facing unknown risks, with marginalized communities potentially bearing a disproportionate burden. Ongoing research into effective solutions for cleaning up contamination and addressing health risks related to PFAS exposure will be vital in the coming years. The call for greater awareness and regulatory action is urgent as the AI industry continues to expand.
