International Law for a Fragile World: Political Risk – Authoritarianism, Information Warfare, and Democratic Fragility (Part II)

International Law for a Fragile World: Political Risk – Authoritarianism, Information Warfare, and Democratic Fragility (Part II)

[Dr Sergey Sayapin is Visiting Principal Research Fellow at the NUS Centre for International Law and Professor of Law at KIMEP University (Almaty, Kazakhstan; on leave).]

Part I of this post argued that contemporary democratic decline proceeds less through dramatic ruptures than through the cumulative, often formally lawful erosion of the institutions that allow political systems to correct themselves, and that international law – organised around discrete rights and flagrant violations – struggles to perceive this process until it is far advanced. This second part examines the dimension of political fragility that has been most profoundly transformed by technology – the information environment – before turning to the paradox that confronts democracies seeking to defend themselves, and to the elements of a legal approach capable of governing political risk before the threshold of irreversibility is crossed.

Information Warfare and the Transformation of Political Vulnerability

Propaganda and political manipulation are not new. States have always sought to influence foreign publics, and political actors have always deployed misleading claims. What has changed is the architecture through which influence operates. Digital platforms permit political communication at unprecedented scale and speed; recommender systems amplify emotionally provocative material irrespective of its accuracy; generative artificial intelligence has collapsed the cost of producing persuasive synthetic text, images, audio, and video; networks of inauthentic accounts allow narratives to be repeated across platforms until they acquire the appearance of consensus; and microtargeting fragments political communication so that different citizens encounter different versions of political reality. Foreign operations need not create domestic divisions – they exploit and amplify those that already exist.

Recent practice illustrates the range of techniques. In September 2024, the United States Department of Justice seized 32 internet domains used by the so-called “Doppelganger” operation, run by Russian companies acting at the direction of the Presidential Administration, which cloned the websites of established news outlets to disseminate fabricated articles and amplified them through influencers and paid social media advertising. Two days before Slovakia’s parliamentary elections in September 2023 – during the legally mandated pre-election moratorium on campaigning – an AI-generated audio recording purporting to capture the leader of a liberal opposition party discussing electoral fraud with a journalist circulated widely, at a moment when neither the media nor the parties concerned could effectively respond. In Moldova, the OSCE/ODIHR observation mission to the 2024 presidential election and constitutional referendum on European integration found that contestants campaigned freely “despite concerns over illicit foreign interference and active disinformation efforts”, and called for stronger action against vote-buying and illicit campaign finance – practices that national authorities attributed to a network directed from Russia. Information manipulation, financial inducements, and the exploitation of distrust are no longer separate techniques but components of integrated campaigns.

The most dramatic legal consequence came in Romania. On 6 December 2024, two days before the presidential run-off, the Constitutional Court annulled the entire electoral process after declassified intelligence reports indicated that a previously marginal candidate had benefited from a coordinated campaign of undisclosed paid promotion and inauthentic amplification on TikTok. Eleven days later, the European Commission opened formal proceedings against TikTok under the Digital Services Act for suspected failures to assess and mitigate systemic risks to electoral integrity. The Romanian case is a watershed precisely because it shows how far legal systems may be driven when manipulation is discovered too late: the only available remedy was to undo an election.

The resulting risk cannot adequately be reduced to “false information”. Democratic politics does not require governments to determine an official truth, and freedom of expression protects ideas that governments consider false, offensive, or politically inconvenient, subject only to restrictions that are provided by law and necessary for a legitimate aim. The systemic problem lies elsewhere – in the deliberate manipulation of the conditions under which citizens form political judgement. Information warfare can succeed without persuading a majority of anything. As Hannah Arendt observed of systematic political lying, a people that can no longer believe anything is deprived not only of its capacity to act but of its capacity to think and judge. RAND’s analysis of the Russian “firehose of falsehood” model points in the same direction: high-volume, multichannel, and internally inconsistent messaging is designed less to persuade than to exhaust. The objective may be not belief but distrust.

This distinction matters legally because international law approaches foreign political interference primarily through sovereignty and the principle of non-intervention. In Nicaragua, the International Court of Justice affirmed that every state is entitled to decide freely the choice of its political, economic, social, and cultural system, and that intervention is wrongful when it uses methods of coercion in regard to such choices. Contemporary information operations sit uneasily with this framework. They may be covert yet non-coercive in the classical sense, working through deception rather than compulsion; they may be conducted through nominally private actors, proxies, influencers, or commercially available tools; and their effects pass through the genuine choices of domestic actors who knowingly or unknowingly reproduce them. Jens David Ohlin has argued that the wrong of covert electoral interference is better understood as a violation of the collective right of a people to self-determination – its right to decide its own political future through its own deliberative processes – than as a violation of territorial sovereignty. That reframing is significant from a risk perspective: it shifts attention from the coercive act to the integrity of the process by which collective choice is formed.

Here the technological and political dimensions of fragility converge. As discussed earlier in this series, technological systems diffuse agency across networks of designers, operators, infrastructures, and adaptive processes. Information operations do the same to political influence. Causation becomes distributed, and attribution arrives late. By the time manipulation is convincingly demonstrated, the election, referendum, or political crisis concerned may already have passed. Political time, like ecological time, can become effectively irreversible.

The Paradox of Democratic Self-Defence

Political risk confronts international law with an especially difficult problem: measures designed to protect democracy can themselves endanger it. The dilemma is old. In 1937 Karl Loewenstein argued that democracies facing fascist movements which exploited democratic freedoms to destroy democracy must become “militant”, and the idea has shaped post-war constitutional orders from Germany’s Basic Law to the European Convention on Human Rights. What is new is that the instruments of democratic self-defence now operate in an information environment in which the line between manipulation and dissent is contested, and in which governments themselves are among the most powerful manipulators.

The same language of security, foreign influence, extremism, and disinformation can be – and is – used to suppress legitimate opposition, stigmatise civil society, and concentrate governmental control over information. Russia’s “foreign agent” legislation, introduced in 2012 under the banner of transparency, became a tool for dismantling independent civil society; in Ecodefence and Others v. Russia (2022) the European Court of Human Rights found that it violated freedom of association. In 2024 Georgia adopted a Law on Transparency of Foreign Influence that the Venice Commission recommended be repealed in its current form, and the European Council subsequently concluded that Georgia’s course had led to a de facto halt in its accession process. At the extreme lies the internet shutdown: Access Now recorded at least 296 shutdowns in 54 countries in 2024, including twelve linked to elections, and at least 313 in 52 countries in 2025. Each shutdown is typically justified as a measure against disinformation or unrest; each also disables the information environment at the moment citizens most need it.

The Romanian annulment illustrates the dilemma within a democratic system. Annulment may have been the only means of preventing a manipulated outcome; it also overrode the votes of millions on the basis of intelligence material that was not fully disclosed. The Venice Commission’s urgent report of January 2025 accepted that constitutional courts may in exceptional circumstances invalidate elections, but insisted that it must remain an ultima ratio, rest on clear evidence of serious irregularities that could have influenced the outcome, be transparent and well reasoned, and be open to effective challenge. The point is not whether Romania’s choice was right but that democratic self-defence must itself be governed by the legality it seeks to protect.

The distinction between democratic resilience and authoritarian insulation is therefore crucial. A resilient democracy is not one from which disruptive information has been eliminated, nor one in which the government possesses sufficient authority to determine what citizens may hear. It is one that maintains the institutional conditions under which citizens can evaluate competing claims and replace those who govern them. Several criteria can help distinguish the two. Legitimate measures target conduct – inauthentic coordination, undisclosed foreign funding, impersonation – rather than viewpoints; they favour transparency over prohibition; they are subject to independent oversight and judicial review; they apply equally to incumbents and challengers; and they are themselves reversible. The legal objective should be the resilience of democratic choice, not the protection of incumbent power.

From Content Regulation to Systemic Risk Regulation

This is why the emerging shift toward systemic risk regulation in the digital sphere is conceptually significant. The European Union’s Digital Services Act requires very large online platforms and search engines to assess and mitigate systemic risks arising from the design and functioning of their services, expressly including negative effects on civic discourse and electoral processes. In March 2024 the Commission issued guidelines on electoral risk mitigation, and the TikTok proceedings opened after Romania represent the first major test of these obligations. The Regulation on the transparency and targeting of political advertising, fully applicable since October 2025, requires labelling and disclosure of sponsorship; the AI Act classifies AI systems intended to influence election outcomes or voting behaviour as high-risk; and the Commission’s European Democracy Shield, presented in November 2025, seeks to integrate these tools with measures supporting independent journalism and civil society. Beyond the Union, the Council of Europe’s Framework Convention on Artificial Intelligence, opened for signature in 2024, obliges parties to adopt measures protecting the integrity, independence, and effectiveness of democratic institutions and processes from risks posed by AI, while the UN Secretary-General’s Global Principles for Information Integrity frame information integrity as a matter of societal trust and resilience rather than state control of content.

These developments remain regionally concentrated, institutionally incomplete, and – in the case of the Union’s instruments – contested by those who see in them a risk of over-removal or extraterritorial overreach. But they reveal an important conceptual evolution. The object of regulation is no longer merely illegal content or an identifiable unlawful communication; it is increasingly the architecture through which systemic political vulnerability is produced. This is the same move from conduct to conditions that this series has traced in the ecological, technological, health, and economic domains.

Democratic Resilience as a Legal Principle

If resilience has normative relevance in those domains, there is little reason to deny it similar importance in the political sphere. It must, however, be defined with care, because political resilience cannot mean institutional survival at any cost. Authoritarian systems can be remarkably durable; stability is not resilience, and governmental continuity is not democratic legitimacy. Democratic resilience should instead describe the capacity of a political system to absorb shocks, resist attempts at domination or manipulation, preserve meaningful pluralism, and correct its own failures without abandoning the rights and procedures that make democratic government legitimate.

So understood, resilience has four interconnected legal dimensions. The first is institutional redundancy: courts, legislatures, electoral authorities, media, civil society, universities, and professional public administrations are not sources of inefficiency but safeguards against concentrated failure, and their independence should be assessed collectively rather than piecemeal. The second is informational integrity without informational control: transparency concerning political advertising, automated amplification, recommender systems, synthetic media, and foreign financing is generally more compatible with freedom of expression than broad prohibitions of vaguely defined “false information”. The third is anticipatory monitoring: the conditions that determine whether an election will be genuinely competitive – media ownership, judicial independence, campaign finance, constituency boundaries, civil society regulation, administrative neutrality – develop years before polling day, so risk assessment must follow trajectories rather than isolated electoral events. The fourth is reversibility: democratic systems become critically fragile when political actors acquire the capacity to make their dominance effectively permanent, whether through entrenched appointments, the abolition of term limits, or the capture of electoral administration. Peaceful alternation in power is therefore more than a democratic convention; it is the clearest evidence of a system’s adaptive capacity.

From Election Protection to Democratic Risk Governance

A risk-oriented approach would not replace existing international rules on elections, political participation, human rights, or non-intervention. It would connect them through a different organising question. Instead of asking only whether democratic rights have already been violated, international law should increasingly ask whether the institutional conditions necessary for their future exercise are being systematically degraded.

This would have concrete consequences. Election observation, which already includes pre-electoral needs assessments, could be integrated systematically with longitudinal assessments of media pluralism, judicial independence, political finance, platform governance, and civic space, so that the cumulative tilting of the playing field is identified as it occurs rather than certified after the fact. International organisations could develop forms of democratic stress-testing, analogous to those used in financial governance, examining how electoral systems, institutions, and information environments would withstand coordinated disinformation, cyber disruption, political violence, or a contested result. Regional regimes could develop graduated responses to systemic erosion – triggered by trajectories rather than ruptures – that engage states well before the point at which suspension or exit becomes the only option.

The language of due diligence, which has run through this series, is equally useful here. States already bear positive obligations to secure the effective enjoyment of political and expressive rights. In an environment where digital infrastructures shape those rights, due diligence should increasingly entail reasonable measures to identify and mitigate foreseeable systemic threats to democratic participation – constrained by legality, necessity, proportionality, independent oversight, and freedom of expression. The aim is not to guarantee particular political outcomes but to maintain the conditions in which political choice remains free, competitive, and reversible.

Platforms cannot be treated as neutral channels. Where a small number of infrastructures organise much of the information environment through which political participation occurs, their design choices acquire constitutional significance. This does not mean transferring censorship powers to private companies. It means recognising that transparency, risk assessment, researcher access to data, accountability for amplification systems, and safeguards concerning political advertising are becoming part of the institutional infrastructure of democratic resilience. Because information operations, political financing, cyber threats, and data markets operate transnationally, international cooperation is indispensable: mechanisms such as the G7 Rapid Response Mechanism on foreign threats to democracy suggest how rapid information-sharing and coordinated attribution might be institutionalised more broadly. The purpose is not to internationalise political choices that belong to citizens but to protect the conditions under which those choices remain genuinely theirs.

Democracy as Global Risk Infrastructure

There is a final reason why political fragility belongs within a broader theory of global risk. Democratic institutions are not merely one vulnerable system among many; they are part of the governance infrastructure through which societies respond to every other vulnerability. A government unable to sustain public trust will struggle to manage a pandemic, as the politicisation of public health measures during COVID-19 demonstrated. A political system captured by short-term incentives will postpone climate adaptation whose benefits accrue beyond the next electoral cycle. A fragmented information environment makes scientific consensus easier to politicise. Economic insecurity deepens polarisation, while polarisation makes redistributive responses to economic shocks harder to sustain. Technological concentration can generate political concentration, while weakened democratic oversight permits further technological concentration.

Risk, in other words, circulates between systems. Part II of the economic post in this series showed that economic fragility has become a principal transmission mechanism for compound crises. Political fragility operates both as an independent risk and as an amplifier of all the others. Democratic resilience therefore cannot remain the specialised concern of constitutional lawyers, election observers, and human rights bodies; it belongs at the centre of the emerging architecture of global risk governance. There is also a deeper normative reason for this conclusion. Risk governance necessarily involves decisions about whose security matters, which vulnerabilities deserve priority, how much uncertainty societies should tolerate, and how the burdens of prevention and adaptation are to be distributed. These are political choices. If the institutions through which they are made cease to be accountable, global risk governance may become more technically sophisticated while becoming less legitimate. Resilience detached from democratic constraint can easily become another name for control.

Governing Political Fragility Before the Threshold

The central lesson of democratic backsliding thus reflects the argument developed throughout this series. International law often becomes most articulate after the decisive threshold has been crossed. It can condemn censorship after a newspaper is closed, arbitrary detention after an opposition leader is imprisoned, fraud after votes are manipulated, and unconstitutional government after institutions have collapsed. Yet political systems, like ecological, financial, technological, and health systems, have thresholds beyond which restoration becomes vastly more difficult – as the experiences of Poland and Hungary show, even when voters succeed in changing course. By the time the violation becomes unmistakable, the institutional capacity to correct it may already have been lost.

The task is not to devise an international legal guarantee that democracies will never deteriorate. No legal order can eliminate political risk, and any attempt to do so would itself threaten democratic freedom. The task is more modest but also more demanding: to identify and protect the institutional conditions that allow democratic systems to absorb disruption without losing their capacity for self-correction. This requires a shift from episodic election protection to continuous democratic risk governance; from exclusive concern with individual violations to the resilience of the institutional ecosystem in which rights are exercised; from post hoc attribution of foreign interference to anticipatory protection of information environments; and from an assumption of institutional stability to the recognition that democracy requires constant maintenance.

International law was designed largely to regulate the conduct of states, but global risk increasingly requires it to govern the resilience of systems. In the political sphere the stakes are higher still, because the system at risk is the very mechanism through which societies decide how every other risk should be governed. Climate systems may cross ecological tipping points, financial systems may experience cascading failure, and technological systems may outpace meaningful human control. Political systems have their own threshold: the point at which institutions formally remain but the possibility of genuine democratic correction has disappeared. A law for a fragile world should not wait until that threshold is crossed. Democratic resilience, in the end, means not the preservation of a particular government or political settlement but the preservation of a society’s capacity to choose another one.

Topics
Featured, General, Public International Law
Tags:
Subscribe
Notify of
guest

0 Comments
Oldest
Newest Most Voted