A group of former Environmental Protection Agency officials has identified at least 30 federal policy changes introduced since the start of 2025 that they argue compromise public health safeguards surrounding the rapid expansion of artificial intelligence data centers across the United States.

The Environmental Protection Network, a nonprofit organization staffed by former EPA employees and administrators, released its findings in a report titled "The Hidden Health Costs of AI Data Centers." According to AI Weekly, the analysis reveals a pattern of regulatory adjustments that either directly facilitate AI infrastructure development or weaken existing environmental protections that apply to such facilities.

Scope and Nature of Rollbacks

Of the 30 documented federal actions, roughly half explicitly reference AI systems or data center operations in their language or intent. The remaining adjustments affect ancillary regulatory domains including energy generation, emissions controls, permitting procedures, and the enforcement mechanisms that federal agencies use to monitor compliance.

  • Seventeen actions directly mention AI or data centers in official documentation
  • Thirteen actions relate to power infrastructure, pollution standards, approval pathways, or regulatory enforcement

Community Health Implications

The Environmental Protection Network's analysis underscores a growing tension between the technology industry's infrastructure demands and the environmental justice concerns raised by communities hosting these facilities. Data centers consume enormous quantities of electricity and water while generating heat, and their siting decisions often disproportionately affect lower-income neighborhoods and rural areas with fewer political resources to contest development.

As AI companies race to build the computational capacity needed for training and deploying large language models and other advanced systems, federal agencies have historically played a critical role in ensuring that expansion does not bypass established health and environmental standards. The new regulatory shifts suggest that calculus may be shifting in favor of faster deployment over stringent oversight.

What's at Stake

The findings raise questions about whether policymakers are adequately weighing the cumulative effects of clustered data center development. Individual facilities may pass environmental review, but regional concentrations can create compounding impacts on air quality, water availability, grid stability, and ecosystem health. A more permissive regulatory environment could accelerate this dynamic without requiring assessment of these broader consequences.

"The Hidden Health Costs of AI Data Centers" calls for renewed scrutiny of how federal agencies balance technological innovation against public welfare protections.

Former EPA staff members argue that many of the rollbacks lack adequate public comment periods and scientific justification. They contend that rushing approvals or weakening pollution standards does not serve the long-term interests of either the AI industry or the communities that bear the environmental burden of its operations.

The report arrives as major technology companies continue announcing expansion plans for data center networks needed to support AI workloads. Federal regulators will face mounting pressure to either re-establish protective guardrails or defend their current approach to the public and Congress.