The doctrine of ‘might makes right’ brings to mind images of generals clothed in military fatigues or the rising tide of strongman politics. The contemporary resurgence of hegemonic masculinity and majoritarianism often focuses attention on coercive power in its most kinetic forms: the border wall, the riot squad, or the violation of territorial sovereignty. Yet, the most profound erosion of legal restraint in the twenty-first century does not arrive with the roar of artillery; it operates silently within the server racks of state and corporate entities. The true modern hegemon is the algorithm.
We have entered an era where technological might is aggressively recoding the very definition of what is legally ‘right’. As states increasingly deploy artificial intelligence for predictive policing – a practice Will Douglas Heaven, writing for the MIT Technology Review has exposed as inherently racist – judicial sentencing, and border enforcement, they execute a subtle but devastating semantic shift. The raw exercise of state power is no longer justified through ideological supremacy or sheer majoritarian force, but through the unassailable veneer of mathematical objectivity. When an algorithmic risk-assessment tool dictates a citizen’s bail eligibility or flags a region for intensified surveillance, the state’s coercive action is sanitised. It ceases to be a political choice and transforms into a computational inevitability.
This is the algorithmic apologia: a framework that legitimises coercion by hiding the structural and majoritarian biases of training data beneath the impenetrable hood of machine learning. The danger lies in how efficiently this mechanism subverts democratic accountability.
In traditional legal frameworks, the exercise of might is met with the friction of due process and the burden of justification. But when ‘might’ is mathematical, its outputs are treated as inherent, objective truths. The state weaponises this opacity, arguing that the algorithm is neutral, and thus, its outcomes are automatically just. By doing so, governments achieve the ultimate authoritarian fantasy: the ability to wield unfettered coercive power while simultaneously claiming the high ground of empirical, unbiased administration. In this new paradigm, might does not merely make right – it computes it.
Evading Appellate Scrutiny: The Black Box Problem
At the heart of the liberal legal order is the principle of contestability. The power of the state is tolerated because it is subject to institutional checks, most notably the mechanism of judicial review. Citizens and their legal representatives possess the constitutional right to challenge arbitrary state action before appellate jurisdictions, demanding that the state articulate the rationale behind its coercion. Yet, the integration of proprietary artificial intelligence into public administration creates a systemic circumvention in this fundamental democratic safeguard.
When governance is outsourced to ‘black-box’ algorithmic models – where, as detailed in the Harvard Law Review editorial board’s analysis of the secret criminal-sentencing algorithm challenged in State v. Loomis (2016) – the predictive logic and training datasets are shielded by corporate trade secrets and intellectual property protections. As such, rigorous judicial scrutiny becomes a structural impossibility. How does a constitutional court review the legality of an administrative decision when the decision-maker itself is an opaque neural network? When challenged, the state cannot produce a transparent and reasoned administrative order; it can only present a statistically generated output – an epistemic trap also reflected in the UK Horizon IT scandal. Between 1990 to 2015, British courts prioritised evidence from faulty accounting software over human testimony, resulting in over 900 wrongful criminal convictions of subpostmasters. This miscarriage of justice puts in stark terms the costs of assuming the infallibility of the machine.
This represents a profound crisis for the rule of law. The cornerstone of a robust legal defence is the ability to cross-examine the evidence and interrogate the logic of the accused. By insulating administrative decisions within proprietary code, the state effectively immunises itself against judicial oversight. It replaces the transparent, albeit flawed, reasoning demanded by administrative law with an unquestionable algorithmic fiat. In this context, ‘might’ manifests as the deliberate obfuscation of process. The state retains its monopoly on coercion but sheds its obligation to provide a legally cognisable justification, rendering the right to appellate redress not just difficult, but conceptually void.
The Geopolitics of ‘Compute’: A New Era of Algorithmic Hegemony
If the domestic weaponisation of artificial intelligence erodes the constitutional mechanisms of the state, its unchecked proliferation on the global stage threatens the very architecture of international law. For over a century, the international legal order has been predicated on the physical realities of sovereignty: territorial integrity, maritime boundaries, and the regulation of kinetic conflict. This approach is reflected most notably in the UN Charter (1945) and the Geneva Conventions (1949). But these texts are structurally incapable of addressing a geopolitical landscape where the ultimate currency of power is ‘compute’ – the advanced semiconductors, massive datasets, and processing infrastructure which powers artificial intelligence.
The pursuit of artificial general intelligence has triggered a great-power scramble for compute, but the theatre of conflict is entirely borderless. We are witnessing the emergence of a bifurcated global hierarchy dictated not by nuclear arsenals, but by algorithmic capability. A concentrated axis of tech-sovereigns – comprising a handful of hyper-resourced nation-states and their deeply entangled transnational corporations – now possesses the unprecedented ability to project power globally without ever crossing a physical border. Through the extraction of global data, and the subsequent exportation of proprietary algorithmic models, these entities impose their own normative standards on the rest of the world.
This dynamic lays bare the current impotence of multilateralism. Traditional international institutions are paralysed, lacking the vocabulary and the jurisdictional reach to regulate a form of power that exists in the cloud rather than on the map. Consequently, nations in the Global South are increasingly reduced to mere data-extraction zones – a dynamic Ruhi Khan analysed in a July 2025 piece for the London School of Economics as the new ‘AI colonialism’ – compelled to rely on algorithmic infrastructures that are fundamentally indifferent to their domestic laws and cultural contexts.
When the rules of global digital commerce, speech, and security are unilaterally hardcoded in Silicon Valley or Shenzhen, international law is rendered a mere spectator. The tech-sovereign operates in a normative vacuum, proving that on the digital frontier, unilateral might dictates the new international right. This was made apparent in 2022, when a corporate actor (SpaceX) unilaterally refused to activate Starlink satellite access over Crimea, dictating military outcomes in Ukraine’s struggle against Russian invasion. In effect, SpaceX bypassed the consensus-based treaty systems that were designed to protect the weak from the strong.
The Right to be Ruled by Law, Not Code
Ultimately, the challenge posed by algorithmic hegemony demands a radical reassessment of how we define and confront power. The danger of the ‘might makes right’ doctrine in the twenty-first century is not merely that the strong will exploit the weak, but that they will do so while convincing the world that their exploitation is a mathematical necessity. To concede that code is law is to surrender the liberal order to a technocratic absolutism, where the fundamental rights of the citizen are subordinated to the relentless pursuit of systemic efficiency.
The defence against this algorithmic apologia requires acknowledging that political action without coercive power may indeed be meaningless, but coercive power unchecked by readable, contestable law is simply tyranny. If international institutions and domestic courts are to survive this technological paradigm shift, they must reject the false sophistication that equates computational capability with administrative justice. The burden must be forcefully shifted back onto the state to prove that their algorithmic tools comply with human rights, rather than forcing citizens to prove that they have been harmed by a black box they are legally forbidden to open.
True strength in the contemporary era is not the unchecked deployment of artificial intelligence. It is the institutional fortitude required to subordinate that technology to human agency and constitutional restraint. Until the international community and domestic legal frameworks adapt to pierce the veil of proprietary code, the algorithm will remain the ultimate arbiter of coercion. As long as machine logic remains immune to human jurisprudence, we are not living under the rule of law; we are merely yielding to the might of the machine.