AI can’t be trustworthy without data center transparency
This article argues that AI cannot be trustworthy if its underlying data center infrastructure is developed through opaque and unaccountable processes. It highlights concerns from communities across Minnesota regarding proposed hyperscale data centers, focusing on issues such as water consumption, electricity demand, environmental impacts, and a lack of transparency in decision-making. The piece advocates for responsible infrastructure governance, including public participation and honest disclosure, as essential for public trust in AI.
An opinion piece published by MinnPost argues that the concept of "trustworthy AI" is undermined by the opaque development processes of the massive data centers that power these systems. Author Eric Ini contends that while technology companies champion responsible AI, communities across Minnesota, including Farmington, Monticello, Hermantown, North Mankato, Eagan, and Worthington, are raising concerns about the lack of transparency, weak accountability, and insufficient public participation in decisions surrounding large-scale data center projects.
The article highlights that community opposition is not merely against technology but stems from governance issues. Residents question whether public officials and developers provide sufficient information on project scope, energy and water demands, and long-term impacts, with reports of nondisclosure agreements further fueling distrust. These concerns, categorized into transparency, accountability, fairness, sustainability, and public trust, mirror the principles found in leading AI governance frameworks like those from UNESCO, NIST, and the EU.
Ini emphasizes that trustworthy AI requires responsible infrastructure governance, including transparent environmental reviews, meaningful public participation, and honest disclosure of impacts. He asserts that public confidence in AI cannot be built solely on ethics statements but must be earned through transparent conduct across the entire AI ecosystem, including its physical infrastructure. The future of AI, he concludes, will be shaped as much by community demands for a voice in infrastructure decisions as by developments in Silicon Valley or international regulatory agencies.