The question of how to regulate artificial intelligence, once framed as a straightforward matter of safeguarding humanity from potential risks, has develop into increasingly complex. The debate, initially centered on establishing limits and accountability for rapidly advancing AI systems, has been reshaped by a recent confrontation between the US government and Anthropic, an AI company positioning itself as a leader in responsible innovation. This clash highlights a fundamental shift: AI is no longer simply a technology to be regulated, but an increasingly essential infrastructure shaping security, economic life, and the very nature of knowledge.
For years, the call for AI regulation stemmed from a growing concern that technological development was outpacing political judgment. The aim was to prevent a future dominated by a handful of companies wielding immense power, and to ensure that technological progress aligned with broader societal values. Still, the recent events surrounding Anthropic demonstrate that the state’s interest in AI extends beyond mere oversight; it encompasses a desire for complete integration and control. This dynamic raises critical questions about the future of innovation, public-private partnerships, and the balance between technological advancement and individual liberties.
The core of the current dispute lies in a January 2026 memorandum issued by US Defence Secretary Pete Hegseth. This directive mandated that all Department of Defence AI contracts include language permitting “any lawful use” of the contracted technologies. This broad authorization directly conflicted with Anthropic’s established usage policies, which explicitly prohibit the deployment of its AI models – most notably Claude – for mass domestic surveillance or the development of lethal autonomous weapons systems. Anthropic, founded in 2021 by former OpenAI researchers, has distinguished itself by prioritizing caution and long-term impact assessment in its approach to AI development.
A Collision of Principles: Hegseth’s Directive and Anthropic’s Restrictions
Anthropic’s commitment to responsible AI has been widely recognized within the tech community. Its “constitutional AI” framework, emphasizing internal constraints and ethical considerations, has served as a model for others seeking to navigate the complex ethical landscape of artificial intelligence. However, the Pentagon’s directive effectively disregards these safeguards, signaling a willingness to prioritize military applications over ethical concerns. Reports suggest that Claude had already been utilized in classified Pentagon operations, including, allegedly, the planning of interventions such as the capture of former Venezuelan president Nicholas Maduro. While these reports remain unconfirmed, they underscore the extent to which the US government is already leveraging AI capabilities.
The clash between Hegseth’s directive and Anthropic’s policies represents a fundamental shift in perspective. What was once lauded as responsible innovation is now viewed, from the perspective of the state, as an impediment to national security interests. The limits Anthropic sought to impose on its technology are now perceived as hesitation in the face of urgency. This highlights a critical point: artificial intelligence is increasingly viewed not as a product to be regulated, but as a foundational infrastructure upon which the ordering of security and economic life depends.
AI as Infrastructure: Redefining Regulation
This shift in understanding necessitates a reevaluation of the very concept of AI regulation. Under these conditions, regulation is no longer simply about limiting harm; it becomes a mechanism for stabilizing the emerging order and directing the distribution of power within it. The current institutional and legal arrangements being developed will shape how technological progress, political responsibility, and authority are understood for years to come. The effects will be determined not by any single dispute, but by the underlying assumptions about risk, security, and public accountability that become embedded within these arrangements.
The conflict with Anthropic vividly illustrates this transformation. The company’s internal restrictions were attempts to preserve zones where technological capability would not automatically translate into application. This presupposes that restraint can be maintained within the innovation process itself. The initial demand for regulation was rooted in the belief that private ambition could not be trusted, that existential risks required external guardrails, and that the state had a responsibility to align technological momentum with human flourishing. Regulation was envisioned as a corrective to potentially catastrophic consequences, tempering raw capability with collective foresight.
The Blurring Lines: Public-Private Collusion
However, the possibility of the US government effectively absorbing companies like Anthropic reveals a deeper dynamic: the close proximity of Silicon Valley and the state. The movement between corporate and governmental authority has become increasingly fluid, with expertise, capital, and strategic capacity circulating across this boundary. What appears as external intervention is often a reorganization from within. This isn’t a simple opposition between technological elites and democratic authority, but an exposure of the degree to which the political system is already structured by technological infrastructures developed by a narrow social stratum.
The state, is not a weaver pulling the strings, but a fly caught in the silk spun by Silicon Valley, its every action dependent on the structure of those threads. The “tech elite” doesn’t merely influence power; it constitutes one of its operating centers. This raises concerns about the concentration of power and the potential for decisions about AI deployment to shape not only outcomes but also the very horizon within which future human decisions are made.
As artificial intelligence becomes more deeply embedded in the fabric of society, it’s crucial to recognize that technology is not merely a collection of tools, but a way in which the world is understood and ordered. The conflict with Anthropic underscores the importance of preserving zones of restraint within innovation, ensuring that technological capability doesn’t automatically translate into application. The next key development to watch will be the Department of Defence’s response to Anthropic’s continued adherence to its usage policies, and whether further action will be taken to compel compliance with the “any lawful use” directive.
This situation demands continued scrutiny and open discussion about the ethical and political implications of artificial intelligence. Share your thoughts and perspectives in the comments below.
Worth a look
