At the end of 2006, TIME magazine made a surprise announcement: ‘For seizing the reins of the global media, for founding and framing the new digital democracy, for working for nothing and beating the pros at their own game’, the headline went, ‘TIME’s Person of the Year for 2006 is You.’[3]

TIME editors had chosen to celebrate not the accomplishments of one individual, but the collective effort of millions of creators of user-generated content on the Internet.[4] It seemed to epitomise Manuel Castells’ network society theorised more than a decade earlier: ‘a social structure made of networks powered by microelectronics-based information and communication technologies’[5] – everyday people working together to create their own rules, norms, and content in an apparently new model of pseudo-anarchic democracy. 

Yet as the ‘Web 2.0’[6] experiment progressed, exclusive online cultures and subcultures were born – some entirely new, others a pastiche of physical world equivalents, but all developing and morphing at unparalleled speed. To the surprise of cyber utopians, many of these are far from the vision of inclusive communities whose thinking broadens through exposure to diverse opinions. Rather, some have developed into closed, self-confirming echo-chambers complete with their own exclusionary language. Challenging exposure to diverse ideas, technology service providers are actively working to tailor services so closely to our tastes that in the near future, even when actively searching for diversity, we may be exposed only to self-confirming content.

Variation 1 suggests that an increasingly limited exposure to ideas contradicting our views is exacerbated by people’s preference for speed over detail in their communications: a barrage of micro-messages – memes, tweets, posts, comments, notifications – containing little extended debate with what they are confronted and further create every day. Human brains can only handle a limited cognitive load,[7] and the onslaught of information, including micro-communications, to which the Internet exposes them increases the need for a filtering system. Yet filtering material according to self-confirming information risks compounding bias, a problem which the frequent use of Artificial Intelligence (AI) and AI-enabled devices in everyday life may amplify.

Variation 1 suggests that TIME Magazine’s celebration of ‘You’ taking the reins of content generation and thereby initiating a much lauded ‘digital democracy’ was pre-emptive. Instead, it outlines a present reality where micro-messages and online echo-chambers restrict space for debate and confirm bias in endless feedback loops, intensified by the growing use of AI-enabled technology. It offers a future where to some extent man is in fact an island, increasingly living in a self-contained information biosphere, at the same time feeding and being fed by self-confirming information.

The ‘You’ creator of the future might yet be a singular and not a plural ‘You’.

                                                                       ***

The Cyber State and its digital villages

In 1996, cyberspace activist Perry Barlow published ‘A Declaration of the Independence of Cyberspace,’ epitomising the cyber-libertarian Zeitgeist. In it, he heralded a new model of society: one without governance or hierarchy, where ‘all may enter without privilege or prejudice accorded by race, economic power, military force, or station of birth… We are creating a world where anyone, anywhere may express his or her beliefs, no matter how singular, without fear of being coerced into silence or conformity.’[8] MIT Media Lab founder Nicholas Negroponte likewise celebrated the potential of the Internet to ‘flatten organizations, globalize society, decentralize control, and help harmonize people,’[9] encapsulating quasi-utopian visions of the ‘Web 2.0’, where ideas would be shared, assumptions challenged, and we would learn from and work with each other within a web-wide ‘virtual community’.[10]

Many consider humans to be a tribal animal.[11] And access to a worldwide virtual network has often precipitated homophily. The Internet enables and encourages people to gather in homogenous groups based on interest, background or outlook, creating separate virtual communities akin to supra-national ‘neo-tribes’.[12] Such communities develop along ideological rather than spatial lines,[13] defined by commonalities rather than diversity in what information scientist Gary Marchionini calls ‘information neighbourhoods’.[14] In fact, people may have a stronger affiliation with such ‘neo-tribes’ than with their physical world equivalents because they are centred on defining characteristics, such as shared beliefs and interests rather than reflecting geographical commonalities.[15] At times, social endorsements within communities can increase exposure to heterogeneous perspectives:[16] in surveys, most social media users report being exposed to a variety of viewpoints.[17] However, research indicates that this is the case primarily in discussions of non-political issues,[18] such as the Olympics or the Super Bowl.[19]  When the identity of a ‘tribe’ is constructed around a more polarising issue, uniformity of opinion can separate people. Within such groups, a polarising effect may be multiplied because information is given and received within a self-confirming echo-chamber,[20] giving a listener a false impression that a particular attitude is more widespread than it actually is.[21]

For some, the digital space offers ‘tribal’ membership to which they may previously have not had access to, were it not for the newly available web connectivity. From mental health support groups to communities actively encouraging suicide; from exclusive online dating forums to paedophilia rings: the best and the worst of humanity finds a place of belonging. While these groups may disperse as quickly as they gather [22] – conjuring up the image of a lava-lamp of online connectivity and communications – it is possible that their very ephemeral nature may serve to harden individual conviction. General Internet use might increase exposure to diverse views. Yet shifting between tribes of the same ‘ilk’ may risk layering different angles of the same viewpoint on top of the last, further entrenching opinions.

Fast talkers

Exclusionary language characteristic of ‘online tribes’ can limit diversity of opinions by occluding outsiders.[23] Elite ICT communities speak ‘Hackish’ (replete with various dialects)[24] and the cryptic question ‘when does the narwhal bacon’ is a phrase specifically designed to screen out non-Reddit readers. A more prolific example is on the online forum 4chan where, as Angela Nagle explains, ‘the constant hazing of n00bs through argot and complex conventions and elite technical knowledge polices the boundaries of the subculture to inoculate it from massification.’[25] Access to and acceptance into such groups seems dependent on whether you ‘think and talk like us’ – if you don’t, eviction is likely.

The effect of linguistic in-groups is magnified by the speed and dynamism of communications often evident within such communities,[26] drowning out dissenting voices. Philosopher Paul Virilio coined the term dromology to explain what moves with speed, quickly comes to dominate territory over anything which moves more slowly.[27] When tweets, posts, comments and memes – each a pastiche of the last – are layered on top of each other with little space or time for debate before the next is added, the result is a cacophony of self-confirming micro-messages moving with such speed that any territory already ceded is quickly reclaimed.

Memes promote unity of thought. Share if you agree

One persuasive case where exclusive language is magnified by speed is the Internet meme, commonly understood as an image combined with a concept or catchphrase. There are informal and usually unspoken norms guiding how some of the Internet’s most prolific memes are used, restricting both understanding and accessibility to particular Internet subcultures.

Take one category of meme: ‘advice animals’ are animal characters accorded particular character traits. A picture of a penguin walking from right to left on a blue background is ‘socially awkward penguin’ (paired with text usually indicating socially embarrassing situations), whereas the reverse image on a red background becomes ‘socially awesome penguin’ (paired with text usually indicating social ‘success’). To confuse the two risks online mockery and ‘social exclusion’ from the relevant online sub-community,[28] in clear contrast to Perry Barlow’s ‘without fear of conformity’ clause.[29]

It is true that certain memes may represent different things for different groups. Pepe the Frog began life innocently as a cartoon in a 2005 comic, before being appropriated in the 2010s by the alt-right movement as a hate symbol,[30] and then re-employed in 2019 by groups of protesters in Hong Kong[31] (apparently unaware of the meme’s connotations in the U.S.) as a progressive symbol of liberty and resistance. While the meaning attributed to this same meme varies between diverse communities, its message is nevertheless clearly defined within the boundaries of each group. Within the confines of each community’s narrative, slightly modified versions of the Pepe meme are repeatedly created and shared, adding layer-upon-layer of micro-communications to crystallise opinion by offering repetitive exposure to the same theme. Indeed, our access to the polyvocality foreseen in Web 2.0 seems often to manifest itself in how many rather than how varied are these voices.

The Daily Me

Limited exposure to contradicting ideas is exacerbated by ever more individual tailoring of online information. Netflix feeds viewers with ‘percentage match scores’, predicting enjoyment of unwatched programmes, thus increasing the likelihood of choosing similar content. Shopping activities affect exposure to adverts, thereby driving consumer behaviour: endless feedback-loops where undertaking an activity may actually make any future divergence from that activity less likely. Perhaps even more concerning is the increased tailoring of accessible information. Pages presented to Facebook users vary depending on their political views, in a self-confirming, political hall of mirrors. Huffington Post and Yahoo News tailor news content directly to readers, giving substance to architect and computer scientist Nicholas Negroponte’s 1995 prediction of a ‘Daily Me’[32] – a virtual newspaper customised for each individual’s taste. Meanwhile, search engines, such as Google, increasingly restrict results according to online behaviour – two people searching for ‘Egypt’ are ever more likely to be presented with entirely different results.[33] This trajectory could signal an Orwellian dystopia where diversified information may not be accessible even if searched for. Indeed, people may even lose awareness that such diversity exists. For many in the future, this ‘Truman Show’ may become a reflection of reality: and why not? If a contrary opinion is googled, results may simply confirm opinion. Activist and entrepreneur Eli Pariser has termed this the ‘filter bubble’[34] – a state of intellectual isolationism which, he argues, foments polarisation across societies and threatens democracy itself:

‘Democracy requires citizens to see things from one another’s point of view, but instead we’re more and more enclosed in our own bubbles. Democracy requires a reliance on shared facts; instead we’re being offered parallel but separate universes.’[35]

Byte-sized information

While online privacy concerns are common, consumers have a vested interest in watching programmes that conform to their individual taste – buying products similar to those with which they have previous experience. Human brains are incapable of absorbing and processing the volume of information already presented to them.[36] This dynamic has only been exacerbated by the influx of information readily available at our fingertips. In a world where over 2.5 quintillion bytes of data are created each day,[37] a filtering system is required. The early days of the Internet anticipated a utopia of information and choice. Yet, as psychologist Barry Schwartz demonstrates, an over-abundance of choice can lead to confusion, disorientation and, ultimately, unhappiness.[38]

The micro-messages described earlier are symptomatic of this reality. Confronted with a complex and often confusing onslaught of information, ideas whittled into digestible bite-size chunks can be welcome and necessary for our ‘cognitive miser’[39] brains which naturally look for shortcuts. Meme creators can compress complex ideas into simple visual packages, while even intricate military and political decisions can be conveyed in a tweet – from declaring kinetic operations[40] to providing legal notification to Congress.[41] At the same time, the speed, lack of detail and sheer volume characteristic of such communications potentially shrinks the time and space for nuanced debate.

As a surreal exacerbation of the issue, a barrage of simplified communications can serve to contribute to the confusion by increasing cognitive load. A review by academics from Oxford, King’s College London, Harvard, and Western Sydney Universities found that ‘the limitless stream of prompts and notifications from the Internet encourages us towards constantly holding a divided attention – which then in turn may decrease capacity for maintaining concentration on a single task.’[42] In a chicken-and-egg scenario from which there seems little escape, the prolific production of micro-communications seems itself to limit the absorption of information in anything other than a micro-message format. 

Thus emerges an ironic reality in which a barrage of micro-messages contributes to an increasingly complex and incomprehensible information environment, thereby requiring information to be increasingly presented in simple, bite-sized slices. Truly nuanced debate becomes less common. This comes at a time when exposure to diversity of thought is already limited by segregating some people into exclusive online ‘tribes’ with the ‘layer effect’ of communications within these groups continually cementing opinions. Underpinning this reality is the dystopian trajectory where access to non-confirming information may be actively restricted. This dynamic not only renders ‘netizens’[43] incapable of finding answers to questions but also potentially unaware of the need to ask questions in the first place.

The AI exacerbator

There is a danger that the growing use of AI-enabled devices in our everyday lives could exacerbate this development in the near future.

The problem of asking machines to think and make decisions like humans, is that the best and worst of human tendencies may be reflected and embedded in this process. AI-enabled personal devices learn and adapt through the information we feed them, tailoring their results accordingly. AI bias based on human prejudice already represents a significant problem: in 2018 Amazon scrapped its AI-based recruiting tool after it was found that female candidates were being screened out. The tool had been trained on historical hiring data and thus ‘learned’ to reflect the bias of a male-dominated tech industry.[44] In the same year, a researcher at the Massachusetts Institute of Technology found that facial recognition software was accurate for dark-skinned women only 35% of the time, compared with 99% of the time for white men[45] (having been exposed to more images of the latter). This trend is particularly problematic in law-enforcement because it can lead to false accusations and arrests.[46]

Today’s communication environment may be set on a trajectory towards a growing number of people being faced with fewer contradicting ideas in their everyday lives. As AI develops, this limitation may increase. If we choose to accept it, the near future could see devices in our personal ‘Internet of Things’ feed us individually tailored reality based on our pre-existing viewpoints and information which they have learned to align with our preferences. This may lead to ever-increasing feedback loops between AI, online communication, and personal preferences, trapping the user in a confirmation biased hall-of-mirrors.

As our use of these devices increases the volume, speed, and complexity of information available to us, we may come to rely on them to distil that information for us to absorb. A barrage of condensed information via micro-messages may ensue, again hampering nuanced debate. Unless a strong effort is made to seek out dissenting opinion, this may take place within an environment where limited exposure to alternative viewpoints restricts our ability to debate and question.

***

This near future vision directly contradicts TIME magazine’s 2006 praise of user-generated content as evidence of ‘community and collaboration on a scale never seen before.’[47] Rather, it is precisely user-generated content that could create information bubbles at an individual level where each person can access different information, yet has neither the tools nor inclination to question it. This could contribute to social fracture by entrenching difference.

In 2006, TIME magazine ended its celebration of ‘You’ with the thought, ‘Web 2.0 is a massive social experiment, and, like any experiment worth trying, it could fail.’[48] What failure means–– just like success––was not defined. However, a future in which it is not the space between us which shrinks, but the spaces we are exposed to as individuals for question, challenge and debate, would probably come quite close.

States have made strong sovereign claims over national cyberspaces over the past thirty years. In doing so, they have driven the breakup of the internet.[49]Just as a shard of glass can break into smaller and smaller fragments, the process of internet fragmentation is continuous and highly irregular.[50] While some states further this division by hardening digital borders, others advance a counter-trend by converging to facilitate greater cooperation.

The debate over global internet governance has been shaped by a tension between traditional, ‘territorially-bound’ understandings of state sovereignty and the internet’s ‘non-territorial’ nature.[51] In the literature on the subject, this debate is equated with a U.S.-China divide – given the United States’ enduring support for a universal internet and China’s fierce advocacy of ‘internet sovereignty’.[52] However, such a dichotomous description fails to capture the complexity of the current internet landscape. While China’s Great Firewall may be the most prominent example of state sovereignty in cyberspace, other states, including Iran and Russia, have also hardened their digital borders.[53] There is even a significant degree of variation in how Western states – particularly the U.S. and EU member states – have approached internet regulation, despite their historic support for a ‘laissez-faire understanding’ of a ‘free and open’, universal internet.[54] Meanwhile, states such as India and Kenya and more across Asia, Africa, and South America are struggling to manage rapidly expanding digital populations.[55] The question for today, then, is no longer whether the internet will fragment, but what this fragmentation will lead to.

The internet today has become a complex, rapidly evolving landscape that has rendered the ‘universal’ internet a relic of the past. Varying approaches to internet governance have emerged across the globe – from the original American model to the ‘full sovereignty’ approach, to a model with regulatory divergences among EU member states, and further to ‘swing states’ that fall somewhere in between. This Variation suggests that states should accept this ‘new normal’ – recognise that the internet has fragmented and the current internet landscape is in a state of constant flux. Rather than decry this fragmentation, states should accept its inevitability and identify opportunities to build a new internet landscape characterised by regulatory blocs. And find a new metaphor to describe the emerging landscape with which we are already living.

The End of the ‘Universal’ Internet

The very concept of fragmentation implies that national cyberspaces are breaking off from a whole, universal internet. However, this inherent assumption of a once ‘universal’ internet is fundamentally flawed. The early internet was ‘universal’ insofar as it was American and Western.[56] As the internet has grown, it has naturally evolved to resemble the realities of the physical world. Rather than mourn this development, states should see a new beginning – embrace a more diverse internet landscape to drive innovation and freedom of choice.[57]

The Original Model

The early internet, developed by the U.S. with support from European partners, was seen as fundamentally free, open, and universal. The military ARPAnet, antecedent to the modern internet, was imbued with Western liberal values from its inception.[58] The U.S. Department of Defense asserted that an ‘open, secure, interoperable, and reliable Internet […] reflect[s] core American values – of freedom of expression and privacy, creativity, opportunity, and innovation.’[59] This early period is referred to as the era of cyber-libertarianism, during which the preference for an unregulated, universal internet reigned supreme.[60]

As the internet grew exponentially, a U.S.-backed ‘multi-stakeholder’ model of internet governance was adopted. It brought to the table non-governmental stakeholders, such as public interest advocacy groups, private businesses, and non-profit organisations, alongside governments.[61] However, multi-stakeholder governance cannot be equated with universal inclusivity. As noted in a UNESCO report on internet freedom: ‘not everyone has decided on everything nor is this what is meant by multi-stakeholder participation’.[62] Power has remained largely in the West and particularly in the U.S., considering that ‘the [internet’s] core technical infrastructure was built in the US and Europe’, and therefore ‘the majority of the twelve organizations that run the internet’s root servers, comprising the core of the internet infrastructure, are American’.[63] It is, therefore, unsurprising that the multi-stakeholder model of internet governance has consistently been criticised by internet stakeholders, including Brazil, India, Russia and China, for its close association with the U.S.[64] The Internet Corporation for Assigned Names and Numbers (ICANN) which manages Internet Protocol (IP) addresses and the Domain Name System (DNS) – both central components to our understanding of the internet – is a U.S.-based non-profit organisation that until 2016 operated under the U.S. Department of Commerce.[65] This relationship reflects the United States’ historic and controversial influence over internet governance.

Although the U.S. is not immune to the global trend toward regulation – a trend propelled by the internet’s encroachment on nearly all aspects of public life[66] – it remains committed to preserving a ‘free and open’ internet through multi-stakeholder governance. The administration of President Donald Trump made clear its conviction that the U.S. should retain stewardship of not only a Western, but an American internet when it declared in an Executive Order that because ‘the internet is a United States invention, it should reflect American values as it continues to transform the future for all nations and all generations’.[67]

Full Sovereignty

America’s heavy-handed influence has led ‘many non-Western nations […][to] perceive the existing internet governance model as a manifestation of American power’.[68] China views the existing internet order as an anarchic space, where a hegemon like the U.S. has exacerbated asymmetry through technological advantage.[69] As internet scholar, Milton Mueller points out, ‘[…] states have for decades if not centuries engaged in power games over resources and strategic advantage and […] view Internet governance from within that framework’.[70] Consequently, to counter the U.S.-backed ‘multi-stakeholder’ model that privileges the American public and private sectors,[71] China, Russia, and Iran have coalesced around a model of full state sovereignty in cyberspace. China and Russia in particular have leveraged multilateral institutions such as the United Nations to advance their shared objective of rebalancing the global internet order.[72]

When China announced its intent to exert sovereign control over its national cyberspace, the notion was ridiculed by then U.S. President Bill Clinton.[73] However, China has remained committed to this principle, gradually achieving full sovereignty over its national cyberspace through a tight censorship net, often referred to as the ‘Great Firewall of China’. The Chinese government justifies rigorous action by highlighting the political implications of internet governance, arguing that ‘cyber sovereignty’ is a question of national security.[74] China is actively looking to establish alliances and pursue its strategic interests in the digital space through efforts to ‘promote its approach to internet regulation abroad, directly challenging the “free and open” philosophy.[75]

One of China’s most significant allies in this arena is Russia. It also aims to advance ‘internet sovereignty’ and undermine U.S.-led promotion of internet freedom.[76] Beyond its longstanding use of censorship,[77] the Snowden scandal and the 2014 Ukraine Crisis have provided the Russian government with ample justification to control its own national cyberspace by creating digital borders within the World Wide Web,[78] described by Russian President Vladimir Putin as a ‘CIA Project’ in 2014.[79]

Another state fixated on internet sovereignty, albeit in a more extreme fashion, is the Islamic Republic of Iran. It set up SHOMA (also known as the 'National Internet', 'National Intranet', or 'Halal Internet') to counter external and internal tensions. The state is better able to throttle foreign connection speeds during politically sensitive periods without crippling critical services.[80] Easier internet access in rural areas is a further benefit.[81] Although debate on its internet governance has lasted for ten years, Iran’s political posture and priorities have shifted remarkably little.[82]

‘Swing States’

If internet governance were represented on a sliding scale, then the multi-stakeholder model backed by the U.S. would stand diametrically opposed to the full sovereignty model. ‘Swing states’, however, would not fit neatly into either camp. Regardless of their form of government, swing states often shift toward the model of internet governance that best serves their domestic priorities at any given time.

Following a debate on internet governance at the World Conference on International Telecommunications in Dubai in 2012, ‘swing states’ have become the focal point of contention.[83] Characterised by ‘mixed political orientation’ and resources to ‘influence the trajectory of an international process’[84] they face a number of challenges, namely protests and disinformation campaigns coupled with a dramatic increase in digital populations. African states, home to ‘the world’s most rapidly expanding digital population’,[85] now favour the sovereign model, leading to an increased government censorship and harsh punishments for content deemed offensive.[86]

Increased state sovereignty over cyberspace has created opportunities for these states to experiment with tightening state control in response to these challenges. States have increasingly implemented internet shutdowns, perhaps the extreme manifestation of governmental control over national cyberspace. Since Egypt first cut off its internet in drastic fashion during the 2011 uprising,[87] the number of internet shutdowns has risen to 75 in 2016 and 196 in 2018.[88] India alone accounted for 134 of those shutdowns, citing problems ranging from students’ cheating during school exams to curbing the spread of information ‘believed to incite communal violence’.[89]

Digital Single Market

The European bloc is typically seen as closer to the U.S. model of internet governance on the sliding scale, largely due to shared values, such as freedom of expression.[90] However, significant fault lines have emerged between the U.S. and EU over regulatory protection of those shared values. Hielke Hijmans, a specialist in EU law and privacy law, recalls this ‘divergence of protective arrangements has been referred to as a transatlantic divide’.[91] This divide manifested in the early 2000s when the American company Yahoo was sued for hosting auctions of Nazi memorabilia on websites accessible in France, and in a momentous decision was ordered to bar French users from accessing such content.[92] Approaches to privacy protection also diverge. The EU’s General Data Protection Regulation (GDPR) states that ‘the protection of natural persons in relation to the processing of personal data is a fundamental right’[93] (L119/1) and consequently implements some of the most stringent protections for online personal data to date. This landmark regulation has helped to secure the EU’s role as a ‘constitutional guardian of privacy and data protection on the internet’ according to Hijmans.[94]

Since then, many EU states have made efforts to advance legislation for the digital domain, touching upon individual rights, civil liberties, and market regulation.[95] However, the legal consensus is that states’ individual efforts to manage their approach are resulting in multi-layered realms of legislation which can ‘counteract and sometimes contradict each other’.[96] Many academics are calling for international frameworks to provide clarity of legislation with regard to trade and economic practices, as well as crime and prosecution, and almost all other aspects of public life.[97] EU member states are modelling avenues of cooperation through measures such as establishing a Digital Single Market (DSM), improving consumer access to goods and services between EU member states,[98] and Schengen-style freedom of movement in the EU region.

***

Less than twenty years after U.S. President Bill Clinton compared regulating the internet to ‘trying to nail Jell-O to the wall’,[99] states around the world have accepted the inevitability of regulation, albeit to varying degrees. China, Russia, and Iran have rejected the original model of internet governance pursued by the U.S. and disagree with the single market project pursued by the EU. At the same time, a great degree of variation exists between oscillating ‘swing states’. These models are not absolute. The fragmentation of the internet is a highly complex and evolving process. Virtual private networks (VPNs) are used by tens of millions of internet consumers in China. They use a VPN to circumvent China’s firewalls and access restricted content.[100] Despite differences and irregularities between and within all models, their existence is proof positive that today’s internet landscape is fragmented.

Nevertheless, fragmentation does not imply the end of a free and open internet; meaning the absence of state-imposed restrictions on internet users’ freedom of expression and access to the World Wide Web. Although observers have often equated a free and open internet with a universal one, the ‘universal internet’ was never truly such. States have always interpreted the meaning of free and open through a lens of self-interest. Germany and the United States both consider their internet spaces to be free and open. Both consistently appear at the top of internet freedom rankings. Yet the instruments and approaches they wield to regulate those freedoms differ significantly.

Eli Noam, Director of the Columbia Institute for Tele-Information, notes that rather than ‘banging our heads against the wall trying to achieve a uniformity that satisfies nobody, we should find arrangements where different systems can coexist’.[101] Today’s internet landscape shows such arrangements are not only possible, but are already in place. Accepting and even embracing internet fragmentation could be the first step toward forming new alliances and shaping the future of the internet on the basis of shared values. A free and open internet can still exist within a fragmented internet landscape, provided that states establish regulatory coherence around the frameworks that define and govern a free and open internet. In this respect, states will have to engage in truly multi-stakeholder and multi-lateral processes. That entails working together and with private tech and social media companies to overcome regulatory differences. In consensus, states may be able to re-shape the future internet landscape, preserving free and open internet space(s) within regulatory blocs that may resemble Schengen-style freedoms and responsibilities.

In the run-up to the UK general election in December 2019, politicians and civil servants warned that foreign actors would attempt to influence the democratic process through covert cyber activity.[102] The UK is the latest addition to a list of democracies facing this threat. Whether in the United States (US), Estonia or Germany, state-sponsored cyber interference has become alarmingly common.[103]

We define cyber interference as state-sponsored online activity that seeks to penetrate a target state’s information environment, with the purpose of influencing internal processes.[104] The open structure of the internet in the West – a defining feature of its communications environment – is vulnerable to abuse by malign actors. Through disseminating disinformation and manipulating social media platforms, cyber interference can polarise or distort political debates, undermine cohesion in target societies, and advance a hostile actor’s strategic objectives. In effect, it is a hybrid threat, operating on the spectrum of legality, clouding accountability and remaining below the threshold of conventional war.[105] 

When can cyber interference be considered a breach of international law? We address this question by examining strategic communications literature and international legal documents, elevating research through interviews conducted with strategic communications practitioners in June 2019.

We argue that under certain circumstances, cyber interference can constitute a breach in international law. Hostile actors may pollute a state’s information environment to such an extent that the target society can no longer tell fact from fiction. By influencing public perceptions, a hostile state may disrupt internal democratic processes in a way that violates the principle of non-intervention.[106]

The growing prevalence, invasiveness, and potency of cyber operations has invited scrutiny by academics, legal scholars, and security agencies seeking to apply international laws of war to the cyber realm. The nature of liberal democracies makes them vulnerable to such actions, due to their inherent pluralism, free media, open economy or system of checks and balances.[107] Such states should recognise the dangers that cyber interference poses to their societal fabric, communicate their position on how it might be regulated, share information and best practices, and create credible mechanisms for attribution and retaliation. By doing so, they may mitigate the risks of cyber interference and adapt the international architecture of norms and values to the threats and technological challenges of the 21st century.

From Cyber Warfare to Twitter Wars

The proliferation of cyberattacks has raised attention to hostile state activity in cyberspace. Operations in the Baltics and viruses like Stuxnet and Petya showed that cyberattacks could qualify as armed attacks under international law, spurring several attempts to apply legal frameworks on conventional warfare to cyberwarfare.[108] Although not universally embraced, the Tallinn Manual, published by the NATO Cooperative Cyber Defense Centre of Excellence, was the first multilateral effort to do so.[109]

However, what can be said for cyber operations that do not cross the threshold of the use of force? Instances, such as the 2015 network hack of the German Bundestag, the 2014 Scottish Independence referendum, the 2016 US presidential election, and the artificial amplification of debates around contentious issues online, are only a few of the numerous cases where such methods have been used.[110] According to data collected by the Center for Strategic and International Studies (CSIS), the number of election-related cyber operations grew from none in 2006 to eight in 2018.[111]

Democracies like the US, India, and Israel were four times more likely to fall victim to cyber-incidents than autocracies like China and Russia.[112] The impact of such incidents can be significant, most clearly demonstrated in instances of cyber interference revolving around national elections. In the 2016 US presidential campaign, Facebook’s General Counsel, Colin Stretch told the Senate Intelligence Committee that accounts linked to the Internet Research Agency (IRA), a well-documented Kremlin-sponsored troll farm, produced over 80,000 election-related posts with a reach of 126 million people.[113] In a race ultimately decided by around 80,000 voters in key battleground states, it is possible that Russian interference influenced the election outcome in some way.[114] If this is indeed the case, then state-sponsored cyber interference is likely to have played a role in determining the political leadership and direction of another state.

The Case for Coercion

There is general agreement in the international community that the basic principles of international law apply to cyberspace, including the principle of non-intervention. Tallinn 2.0, a follow-up to the Tallinn Manual that explores the relationship between international law and cyber operations beyond the dimension of cyberwarfare, argues that respect for sovereignty covers the closely related principles of the prohibition of the use of force and non-intervention in the internal affairs of other states.[115] A 2015 report by the UN Group of Governmental Experts (UN GGE) on risks arising from states’ use of ICTs echoes this, noting that states must honour their commitments to the UN Charter and other principles of international law, including non-intervention.[116] While the UN GGE was dissolved in 2017 after failing to reach consensus, subsequent work from other organisations, including the European Commission´s 2018 report on tackling online disinformation, demonstrates a sustained trend in this direction.[117]

The principle of non-intervention is a foundational norm in international law. The 1933 Montevideo Convention on the Rights and Duties of States holds that ‘no state has the right to intervene in the internal or external affairs of another’.[118] This is reflected in the UN Charter’s articles on sovereign equality and the prohibition on the threat or use of force, which protect a state’s territorial integrity and political independence.[119] A more forceful iteration of the principle of non-intervention is found in the 1970 Friendly Relations Declaration, which states that breaching this principle constitutes a violation of international law: ‘Every State has an inalienable right to choose its political, economic, social and cultural systems, without interference in any form by another state’.[120]

Cyber-interference may be considered to fit comfortably within the realm of unlawful interventions. However, the 1986 International Court of Justice (ICJ) ruling regarding the illegal US intervention in Nicaragua complicates matters. The Court held that intervention may only be deemed unlawful when it uses methods of coercion.[121] Informed by the time and context of its writing, the ruling offers a narrow interpretation of coercion limited to forcible measures. Tallinn 2.0 concurs with the ruling and argues that without coercion, interventions cannot be counted as unlawful.[122]

However, closer inspection reveals gaps in the Court’s interpretation of coercion. Tallinn 2.0 argues that coercion can be any act ‘designed to deprive another State of its freedom of choice’ or to influence outcomes within its domaine réservé – matters on which it is within a state’s sovereign rights to decide freely.[123] The most important of these were agreed to be the choice of political system and its organisation which ‘lie at the heart of sovereignty’.[124] In an early criticism of the ICJ’s ruling, Professor Lori Damrosch of Columbia University urged the international legal community pay close attention to non-forcible measures, such as foreign funding for political campaigns and propaganda.[125]

Furthermore, both the Court and Tallinn 2.0 concur that such acts can be either direct or indirect. Tallinn 2.0 offers an illustration of an indirect cyber operation of this nature: Imagine State A releases sensitive information online regarding State B that brings about a political crisis in State B, eventually leading to policy changes that would not have happened otherwise. Most experts behind Tallinn 2.0 agreed that this would be an example of unlawful intervention.[126] In effect, this conclusion casts a new light on incidents like the DNC hack in 2016 and bolsters the view that cyber interference can be considered coercive.

The implications of this conclusion pave the way for reevaluating the nature and implications of coercion in the 21st century. Professor Steven Wheatley purports that cyber interference, in the form of a sustained campaign of disinformation, should also be considered coercive – and therefore illegal – when it ‘”pollutes” the democratic debate […] to such an extent that it undermines the capacity of the population to make a meaningful choice based on the facts…’.[127] According to Wheatley, this effectively coerces the population to choose a course of action to which it did not knowingly consent, and hence denies people their right to political self-determination.[128] Professor Damrosch introduces a human rights element, purporting that impinging on the right to self-determination violates both the principle of state sovereignty and citizens’ rights to self-governance.[129]

Consequently, there seems sufficient legal grounding to consider certain instances of cyber interference as internationally wrongful acts when these amount to coercive behaviour aimed at intervening in another state’s domaine réservé. The question now is: what can be done about it?

Putting Words into (State) Practice

First, states that are vulnerable to cyber interference must work to entrench and reinforce the notion that it is considered illegal. To shape nascent international norms, states must contribute to state practice and opinio juris (belief in a legal obligation) in a coordinated manner. Currently, few states condemn instances of cyber interference as internationally wrongful acts. In reaction to alleged Russian interference in the US presidential campaign, President Obama initially responded with whispered warnings to President Putin.[130] This approach may derive from the difficulty of determining attribution or from the fact that many states engage in cyber interference themselves.[131] However, due to the risks posed by cyber interference to state sovereignty, states should set an example in their domestic behaviour, avoid coercive cyber interference, and clearly communicate their position in multilateral platforms, such as the UN’s Open-Ended Working Group (OEWG).[132]

In addition to publicly condemning cyber interference, states may consider means of retaliation. US criminal investigations and economic sanctions against Russia could provide a basis for state practice. But such reactions must be more consistent and widespread to be considered evidence of an emerging norm. Ironically, furious Russian denials of conducting cyber interference, despite overwhelming evidence to the contrary, actually support the existence of opinio juris against committing cyber interference – had it been legal, there would have been no difficulty in admitting it.[133]

Second, states should incline to build trust and transparency through sharing information and best practices. Currently, states deploy different methods to strengthen the resilience of the various potential targets of cyber interference. Sweden focuses on defending its electoral process[134], Singapore promotes counter-disinformation legislation[135], India pressures social media companies to install technical restrictions on popular platforms[136], and the Baltic states emphasise accountability for perpetrators.[137]

Even more importantly, states must willingly disclose instances of when whether they are dealing with cyber interference at all. Driven by the fear of revealing their own cyber tools or vulnerabilities, states have been tight-lipped about their experiences in this regard.[138] However, by disclosing some of this information within the confines of national security considerations, states will both learn how to boost their defensive capabilities and strengthen their legal case by informing international legal experts on emerging state practice.

Finally, states should consider creating international mechanisms of attribution. Tallinn 2.0 stipulates that customary international law regarding state responsibility applies to activities in the cyber domain. Consequently, a state would bear international responsibility for unlawful cyber interference credibly attributed to it, either directly or through a non-state actor.[139] Attribution is critical to justifying retaliatory measures. However, as noted by General Dan Efrony and Professor Yuval Shany, anonymity afforded by cyberspace and the lack of widely accepted mechanisms of attribution make timely and reliable claims of attribution difficult to establish.[140] In spite of these difficulties, however, there is a record of several incidents of cyber interference that have been successfully detected and attributed. Advances in data science and artificial intelligence will assist in quickly gaining reliable intelligence, allowing for the attribution of blame on public and diplomatic levels.[141]

***

The speed of technological progress in recent decades has challenged long-held truths regarding the norms and values that have traditionally underpinned the international system. The principle of non-intervention in another state’s internal affairs is the bedrock of state sovereignty. In the past, violations of sovereignty could only be feasibly committed via a limited set of, usually coercive, means. Today, ICTs have opened myriad opportunities for states to undermine adversaries without firing a shot.

Interested states must communicate their position on the impermissibility of cyber interference to adapt the frameworks and instruments international law to the challenges of the 21st century. This can be achieved by condemning cyber interference as illegal when it amounts to coercive behaviour and threatens the freedom of their population to determine their own fate.

Once states speak with a united voice, there is also potential for collective action. Sharing information and best practices should inform domestic and international efforts to counter cyber interference and hold perpetrators accountable. By grasping how communications platforms can be used to shape political debates and undermine political establishments, strategic communicators can affect the rules binding those platforms. Evidently, cyber interference will always persist to some extent as a feature of interstate conduct. Yet experience shows that concerted global efforts can effectively curb aggressive state behaviour. We propose that the same can be done with cyber interference, should states use their words and actions wisely.