The Biden Administration has issued two significant executive orders this week that signal an important evolution in how the U.S. government approaches artificial intelligence. Coming in the final days of the administration, these orders represent a capstone to Biden's technology policy.
While much attention has focused on AI regulation and risk management, these orders—including a January 16th order on innovation in cybersecurity and a January 14th order on AI infrastructure—represent something different: recognition that AI capability is becoming critical infrastructure for effective governance.
However, the framing in national security terms—with a focus on ensuring competitiveness with China and advancing economic competitiveness globally—missed an opportunity to bolster technology’s role in strengthening democratic institutions, improving governance and public service delivery.
From Control to Capability
The first order, focused on AI infrastructure and clean energy, and the second, on cybersecurity innovation, moves beyond asking "how do we control AI?" to address "how do we build and use AI?" Rather than treating AI primarily as a technology to be regulated, these orders position AI as both essential infrastructure and a vital tool for government operations, a significant shift.
The infrastructure order seeks to advance administration priorities around AI advancement without comprising values of climate action and worker protection. By establishing a framework that balances technological advancement with environmental protection, worker rights, and community interests, these orders could serve as a model for how democracies can develop AI infrastructure in a way that strengthens public institutions while preserving critical social and environmental priorities.
The order lays out an ambitious approach that brings together multiple federal agencies. For example, the Departments of Defense and Energy must identify suitable federal sites for AI infrastructure by early 2025, while agencies have 180 days to identify best practices for "public participation and governmental engagement in the development of potential siting and energy-related infrastructure for data centers, to include practices for seeking input on potential health, safety, and environmental impacts and mitigation measures for nearby communities." The order specifically directs the exploration of nuclear power, geothermal energy, and other clean energy sources on federal lands to power these AI facilities.
The order makes it explicit that private industry, not taxpayers, must bear the cost of development of the energy needed to power AI innovation: "The companies developing, commercializing, and deploying AI must finance the cost of building the infrastructure needed for AI operations, including the development of next-generation power infrastructure built for these operations." This principle extends to protecting consumer interests, with specific requirements that AI infrastructure development must not raise energy costs for American consumers and businesses.
Building Capability Through Technology and Operations
What makes these orders significant is their recognition that the government must be both an enabler and user of AI. While the infrastructure order establishes frameworks for developing AI compute capacity on federal lands, including potential nuclear power deployment, the cybersecurity order demonstrates how AI can enhance government effectiveness. It lays out concrete operational improvements like automated vulnerability detection and patch management across federal systems, showing how AI can strengthen basic government operations.
Beyond National Security
While these orders move beyond treating AI primarily as a risk to be managed and instead recognize AI as essential infrastructure for effective governance and economic vitality in the 21st century, they frame AI as a geopolitical arms race.
The defense and national security framing -- and the exclusive focus on private sector AI development--risks obscuring the crucial importance of AI for improving governance. The real promise of AI use, however, lies not in winning a technological arms race, but in demonstrating how democratic societies can harness AI to better serve their citizens.
