By Nai Lee Kalema
BLOG series:
Notes from the New Frontier of Power
On January 20, 2025, President Trump issued the to revoke those of the previous administration in what it explained was an effort to repair and restore common sense to the US Federal Government. It stated that “the into our institutions has corrupted them by replacing hard work, merit, and equality with a divisive and dangerous preferential hierarchy,” and, “Climate extremism has with regulation.”
Trump also issued the , pending review, and Executive Orders, imposing a federal-government-wide hiring freeze (except for national security, immigration enforcement, and public safety positions) and an immediate return to full-time office work for all federal workers.
In a further move, he ordered the Office of Management and Budget (OMB), Office of Personnel Management (OPM), and DOGE directors to submit plans within the next 90 days to Finally, President Trump issued several Executive Orders withdrawing the U.S. from the World Health Organization (WHO), pausing any transfer of U.S. funds or resources to the organization, and halting any pacts, accords, and commitments it made under the .
The U.S. Department of Government Efficiency (DOGE)
On January 20, 2025, President Trump issued an establishing and implementing the Department of Government Efficiency (DOGE). This organization aims to modernize federal technology and software to maximize the efficiency and productivity of government-wide software, network infrastructure, and information technology systems by increasing data integrity, improving data collection and synchronization, and enhancing interoperability between agency networks and systems. Under that Executive Order, the United States Digital Service was reorganized and renamed the United States DOGE Service (USDS). This body, led by the USDS administrator, was established as a temporary organization, U.S. DOGE Service Temporary Organization, designed to advance the President’s 18-month DOGE agenda before terminating on July 4, 2026. In consultation with the USDS administrator, each U.S. agency head is required to establish its own DOGE team, typically consisting of at least four employees (i.e., a DOGE team lead, an engineer, a human resources specialist, and an attorney) to coordinate their work with the USDS in implementing the President‘s DOGE agenda.
The chief enabler of DOGE’s of the U.S. Federal Government initiative is artificial intelligence, accelerating what some scholars have referred to as “” “,” “,” “,” and “.” In other words, algorithmic governance. In light of this upcoming momentous reorganization and shift in the federal government, let’s explore what this massive acceleration of algorithmic governance could mean for democracies going forward.
Algorithmic Governance as Capture
Central to this particular approach to public-sector digital transformation are data, AI, and digital infrastructures. As governments’ reliance on data and data-intensive technologies has grown, their capacity to govern and make decisions concerning administration, finance, policy, and regulations has become dependent on their policy elites’ capacity to manage their political–economic relationships with the digital sector (ref. Margetts & Dunleavy).
With governments ever more reliant on the digital platforms of Silicon Valley companies and cloud and data-intensive infrastructures for their core day-to-day functions, Silicon Valley firms have been able to reshape governments more actively, for example, by strategically reducing the presence of existing technological incumbents (e.g., IBM) in public systems over time through cultural, organizational, and technological developments that have transformed the political-economic relationships between states and big-data companies (ref. Margetts & Dunleavy). Core to algorithmic governance is data extraction from the ever-expanding public datasets, which are a byproduct of the hyperscaling of digitalized public services, systems, and bureaucracies. Such data extraction serves as another and value capture by tech giants.
Further, the algorithmization of governments is leading to their “DCaaSization” (data center as a service) as they increasingly migrate and infrastructure their data to adhere to the epistemic structures of corporate clouds, leaving the AI in public systems locked behind corporate terms and increasing informational asymmetries between governments and . Governments’ outsourcing of their core information-processing capabilities to corporate technological service providers is divesting critical capabilities and skills from states, turning the state from “a technological administrator to a tenant of external vendors and .” Further, governments’ outsourcing of their critical informational processing has diminished the institutional competency and oversight of the technical systems and processes of some public organizations due to their increased reliance on private companies. This has led to the hollowing out of states, turning them into the “” of cloud empires.
What Has Algorithmic Governance Meant for Democratic Governments So Far?
Platforming Governments
Prominent technologists have explored algorithmic governance through the government as a platform metaphor, referring to governments’ creation of interoperable measures, standards, technologies, code bases, digital stacks, open datasets, and open-source software to provide the private sector with a foundation for creating new services and products, with digital infrastructure operating as the . The government-as-a-platform model of algorithmic governance is emblematic of a broader public administration paradigmatic shift that presents the digital platformization of governments as a means for more public value creation, which according to O’Reilly would “encourage the private sector to build applications that government didn’t consider or doesn’t have the resources to create.” O’Reilly provides a hypothetical scenario of how more public value could be created in the health context, explaining how a Medicare rate-setting algorithm could be used with real-time open government data to incentivize more efficient public health spending by linking public health insurance reimbursements to performance objectives. However, in practice, such approaches have had decidedly mixed and even harmful results.
For example, an AI tool was allegedly used to wrongfully deny rehabilitative care services to older and , contravening their . In October 2023, class-action lawsuits were filed against UnitedHealthcare and Humana, two of the U.S.’s largest health insurance companies, for their alleged illegal and inappropriate use of a proprietary AI tool by NaviHealth (a UnitedHealthcare subsidiary). This tool created an algorithmic risk score to determine payment cutoff dates, overriding clinical decisions made by doctors and resulting in inappropriate rehabilitation care coverage denials for . Moreover, this issue was alleged to have been to on these plans.
Merging Tech and State
Amid the emergence of algorithmic governance, we are experiencing the merging of tech and state power similar to that already taking place in many other countries, with the strategic interests of tech giants becoming ever more . For example, we are witnessing the systematic dismantlement of academic research by labs and civil society institutions on algorithmic manipulation and platform accountability (e.g., Musk has called for researchers on election interference and misinformation to be prosecuted, and Trump has called for universities that perform such research to have their non-profit status suspended). Further, as part of U.S. DOGE’s downsizing efforts, AI may be introduced as a replacement for up to 75% of the federal workforce rather than as a complementary tool subject to oversight. Finally, the U.S. is cultivating a de facto regulatory vacuum by wielding bilateral trade agreements to override other countries’ enforcement of their democratically derived digital regulations and policies against U.S. tech giants (e.g., enforcement of the EU’s Digital Services Act against X).
What Might Algorithmic Governance Mean for Democratic Governments in the Future?
At the more extreme end, prominent Silicon Valley actors have proposed disrupting democracies altogether through the advent of network states. Referring to and instead choosing to rename it “AI bias,” Balaji Srinivasan—co-founder of Counsyl, ex-CTO of Coinbase, and former general partner at the venture capital firm Andreessen Horowitz—came up with this concept, referring to geographically decentralized, digitally interconnected, and real-time algorithmically governed societies that eventually obtain physical territories to replace nation-states altogether. These network states would be led by a “Network Leviathan” of “grays”—atheist and anti-statist technological progressives. Moving from theory to practice, the network state is being tested in real life. For instance, Afropolitan, an employing a community-as-a-service business model, raised $2.1 million in venture capital from Srinivasan and over 25 other investors to create the world’s first-ever internet country.
Democracy and AI technology are neither sufficient nor necessary conditions for each other, meaning that AI’s ability to . What the U.S. DOGE and, more broadly, algorithmic governance may hold for the future of the U.S. government and democratic governance is not yet known. If history is a teacher, what remains clear is that any action, such as algorithmic governance, that diminishes human agency, corporate accountability, and regulatory public oversight while increasing democratic deficits by design is unlikely to bode well for the future of American democracy.
