Introduction

Abstract

Narratives provide the storylines of conflict and in doing so become an arena of conflict themselves. When states mount information campaigns against each other, they are trying to change the narrative. The digital platforms of the new information environment have been identified by various analysts as a significant factor in contemporary strategy and crisis management. But while social media is noisier and more chaotic than traditional media, and unprecedented in its immediacy and accessibility, has it thus far been a game changer in strategic affairs?

In this Adelphi book, Sir Lawrence Freedman and Heather Williams examine the impact of state-led digital information - or disinformation - campaigns in four contexts: the India-Pakistan crisis over Kashmir in 2019; the heightened tensions between the United States and Iran following the assassination of Qasem Soleimani in 2020; China's messaging in response to the COVID-19 pandemic from 2020–22; and the Russia-Ukraine crisis from 2013–23. While noting the meaningful consequences of digital information campaigns, in each case the authors call for a sense of perspective. Such campaigns are only one aspect of wider political struggles. They are also difficult for their initiators to control, and less likely to influence foreign audiences than domestic ones. Overall, the authors argue, there is little evidence so far to suggest such campaigns will have as much influence over contemporary crises as the classical instruments of military and economic power.

This book assesses the impact of state-led information campaigns, largely conducted through social media, on crisis escalation and the international competition for strategic advantage. By ‘information campaign’ or ‘information opera tion’ we mean an attempt by an individual or a group – which could be a state – to establish, shape or challenge a narrative; this effort may encompass what could be described as ‘propa ganda’, ‘public diplomacy’ or ‘disinformation’, or even ‘fake news’. By a ‘narrative’ we mean a storyline that interprets events in a way that legitimises the narrator’s behaviour and encourages others to act in a way that serves their purposes. There is nothing new about such activities, but this book seeks to explore their significance within a contemporary information environment that is heavily influenced by the internet and social-media platforms.¹

Given the speed with which information can spread on social networks, the relatively low cost of its propagation, and the potential consequences of national narratives being shaped by these means, various analysts have identified these networks as highly significant factors in contemporary strategy and crisis management. We agree that digital information campaigns have potential strategic significance and deserve careful study, but having examined a number of information campaigns during recent international crises, we are also cautious in assessing the effects of such campaigns. They need to be located within their broader context. Whatever their instigators may intend, their actual effects can be hard to predict or control. Public perceptions turn out to be quite difficult to change and not easily shaped by information from unfamiliar sources. This is why information campaigns are often more effective with domestic audiences, to whom they are more naturally tailored, than with foreign ones. With some exceptions, narratives do not travel well. Influencing thinking in other countries usually depends on knowledge of pre-existing divisions that can be exploited. At times of crisis, social media can also be used to transmit messages between political leaders in an unusually direct and explicit form. Our analysis is concerned with the extent to which either these broader information campaigns or more elite-level messaging can affect crisis behaviour, including the potential for escalation.

The new information environment

Social-media platforms such as Twitter, Facebook, Instagram, WhatsApp, VKontakte and WeChat have become the media sources of choice for many users around the world, including leading political figures. Less than 20 years ago, these platforms seemed to be a revolutionary development; now they are accepted as one of the primary means by which individuals, organisations and governments engage with their discrete and wider communities. Although our focus is predominantly on Twitter, due to its status as the pre-eminent platform for official statements and news in Western countries, we also examine other platforms as their role and significance varies by country. (We are aware that recent problems with Twitter have led to the development of alternative platforms performing a similar function, such as Bluesky and Threads. To add to the confusion, Twitter has been rebranded as ‘X’.)

Marshall McLuhan famously observed that ‘the medium is the message’, referring to the way that forms of messaging affect social structures irrespective of content.² A post on Twitter (a tweet), for example, is shaped by its sudden appearance and its brevity – the first tweets were only 140 characters long, although this was later doubled to 280. Like other social networks, Twitter is a ‘noisy’ medium, so a particular message must compete for attention with a multitude of others being posted simultaneously. It is no simple task to cut through and gain notice. Every message is at risk of being crowded out as thousands more soon pile up behind it. The digital age is marked by a surfeit rather than a scarcity of information. Individuals can choose between a multitude of alternative sources of news and stories to consider, rather than making do with a few easily accessible sources. Without much effort it is possible to scan numerous sites for information, detecting material that would never have been found in earlier times. Much of this information appears in largely headline form, not requiring careful comprehension, with little or no sense of the reliability of the source. Because individuals can choose whom to follow on social media, they can pre-select their information, which often aligns with their pre-existing views. So, despite the mass of diverse material available, often the effect is most powerful in reinforcing existing attitudes.

The billions of messages posted worldwide each day support a variety of social and political functions, benign as much as malign. Many are about everyday matters such as holidays, films, investments and charitable giving, rather than political views and partisan commitments. Even messages on more contentious topics tend to be aimed less at changing views than at demonstrating affinity with those who already share those views. Attention has naturally focused on the darker side of social media, especially the ease with which provocative, inflammatory and defamatory material can be generated and spread. Such material is frequently personalised and directed at celebrities and politicians or used to target specific groups. Users can quickly and simply enter into another country’s debates anonymously or under an assumed name without any requirement to be open or honest about their identity.

Social-media platforms are subject to the policies of their owners and government regulators. The ability of a few companies or even individuals to control how people engage with each other raises questions of unaccountable power. This also relates to companies’ ability to learn about their users’ lives and to exploit this information for targeted advertising, or to sell it to organisers of political campaigns. As a form of communication, social networks continue to evolve.³ As instruments of international communication, they can facilitate frank negotiations of an arms-control agreement, while they can also be manipulated to stir up nationalism during a crisis.⁴ Social media also plays a role in contemporary geopolitics and conflict: world leaders have used it to threaten military strikes, call for peace, open negotiations and appeal to their political bases.⁵

The Trump effect

The question of social media’s role in advancing political narratives gained in salience through Donald Trump’s use of Twitter. He was an early adopter of the platform when his political ambitions were nascent; he later made it his default mode of communication as a US presidential candidate and then as president. His example first raised concerns about how Twitter activity might lead to instability at home and abroad. His account was suspended in early 2021 because of his false allegations that the November 2020 presidential elections had been fraudulent and the perceived risk, in the aftermath of the attack on the Capitol on 6 January 2021, that his Twitter account could incite further violence.⁶ The suspension was lifted after Elon Musk acquired Twitter and became its CEO in October 2022.

After he opened his Twitter account in 2009, Trump’s early tweets were bland, but from 2011 onwards he began to use them to augment and adapt his brand as a combative, outspoken personality. Soon president Barack Obama became a major target as Trump moved further into the political sphere. By the time he began his campaign for the Republican presidential nomination in 2015, he had a mass following and understood more than rival candidates about how social media could be used for political advantage. Trump pioneered a Twitter-based demagoguery, exploiting the polarisation the format encouraged, and continued this into his presidency. His tweeting then became even more significant, not only because of the inherent status of his office but also because he could make his views known while bypassing the normal bureaucratic checks. Before his account was suspended on 8 January 2021, Trump sent 26,236 tweets as US president.⁷ Irrespective of their appropriateness or literacy, most could be understood as authoritative presidential statements. Trump set an online path for others to follow, including Indian Prime Minister Narendra Modi and Ayatollah Ali Khamenei of Iran.

As head of state, Trump’s tweets were distinctive because they were cast as informal interventions. In principle this was not significantly different from unscripted or poorly expressed remarks at a press conference. Yet Trump’s tweets were more deliberate and could also exploit Twitter’s special reach and immediacy. Their potential importance first became clear in the case of North Korea. On 2 January 2017, shortly before his inauguration, Trump issued his first Twitter warning to the hermit kingdom:

North Korea just stated that it is in the final stages of developing a nuclear weapon capable of reaching parts of the U.S. It won’t happen!⁸

To Trump’s frustration, this did not stop Kim Jong-un from testing nuclear weapons and missiles. In July 2017, when a North Korean missile landed close to Japan, Trump responded on Twitter:

North Korea has just launched another missile. Does this guy have anything better to do with his life?⁹

As the situation deteriorated, with reports that North Korea could mount a small nuclear warhead on a missile, Trump responded not with a tweet but with remarks to journalists delivered from one of his golf courses: ‘North Korea best not make any more threats to the United States. They will be met with fire and fury like the world has never seen.’¹⁰ This was followed by a speech at the United Nations during which he described the North Korean leader as ‘Little Rocket Man’ – a term he had also used on Twitter – using a reference to an Elton John song that included a dig at Kim’s height. Kim responded by promising to ‘definitely tame the mentally deranged US dotard with fire’.¹¹

Trump decided that Kim was a bully. Bob Woodward reports that he told an aide, ‘He’s a tough guy. The way to deal with those people is by being tough. And I’m going to intimidate him and I’m going to outfox him.’¹² After Kim’s New Year’s Day 2018 address, in which he expressed confidence that the United States was deterred because ‘a nuclear button is always on the desk of my office’, Trump retorted on Twitter:

North Korean leader Kim Jong Un just stated that the ‘Nuclear Button is on his desk at all times.’ Will someone from his depleted and food starved regime please inform him that I too have a Nuclear Button, but it is a much bigger & more powerful one than his, and my Button works!¹³

The most dangerous moment came when Trump planned to tweet that he was ordering the departure from South Korea of all US military dependants – the thousands of family members of the 28,500 troops stationed there.¹⁴ This was likely to be taken as a sign of imminent attack by the US armed forces. He was dissuaded from posting the tweet. Yet when in March 2018 the South Koreans conveyed an invitation from Kim to a summit, Trump accepted immediately. The leaders met in Singapore in June that year and gave the impression of cooperation and diplomatic progress, even to the point of denying there was an issue when it became clear that Kim had no interest in the ‘denuclearisation’ process that was supposed to be at the heart of the summit communiqué.

In Jeffrey Lewis’s clever fictional novel The 2020 Commission, a report examines the reasons for a North Korean nuclear attack on the US in 2020, with a presidential tweet playing a crucial role in the story.¹⁵ The novel depicts the brief rapprochement of 2018 followed by a deterioration in relations, marked by a series of hostile and mocking tweets from Trump as North Korea fails to denuclearise. A crisis arises when North Korean air defences shoot down a South Korean civilian airliner that has drifted off course, having mistaken it for a military plane. Without consulting Washington, South Korean president Moon Jae-in retaliates with a small number of conventional-missile strikes against North Korean targets. Kim misinterprets a tweet from Trump boasting that ‘LITTLE ROCKET MAN WON’T BE BOTHERING US MUCH LONGER’ and mistakenly assumes an attempt to destroy the North Korean regime is under way. He responds with nuclear strikes against South Korea, Japan, and US bases in the region.

Fiction can often illuminate real strategic issues. In principle, this scenario demonstrates the power of social media in the wrong hands. But it also reminds us that the effect of such messages depends on context. In the story, Trump’s tweet was not an isolated event, and it was lent authenticity by Trump having been baiting Kim for months. It was also posted without deliberation or advice, and in ignorance of the actual state of affairs. This might have been a feature of Trump’s tweeting in reality, but it is a practice that others are unlikely to follow. The issue in this fictional account was not so much the use of Twitter but the ability of a leader to make an inflammatory statement without heeding advice or having their bureaucracy prepare for the consequences.

In an example of an actual tweet with an unintentionally escalatory effect, in August 2018 Canadian foreign minister Chrystia Freeland responded to the arrest of Saudi humanrights activist Samar Badawi:

Very alarmed to learn that Samar Badawi, Raif Badawi’s sister, has been imprisoned in Saudi Arabia. Canada stands together with the Badawi family in this difficult time, and we continue to strongly call for the release of both Raif and Samar Badawi.¹⁶

The tweet, aimed at a Canadian domestic audience, had international repercussions because Riyadh considered it insulting and provocative. It provoked a furious Saudi response that included expelling the Canadian ambassador, threatening to freeze all new trade and investment, halting flights between the countries, and recalling Saudi students studying in Canada. Iyad Madani, a former Saudi information minister, wrote that Canada was interfering in his country’s domestic affairs: ‘Canada blundered because it seems to have ignored and forgotten that civil society and political social development are best left to the dynamics of each society.’¹⁷ Ultimately, Canada did not back down and not all the Saudi threats were implemented.

Twitter in context

While Twitter remains the social-media platform most used for political statements and for individuals to access news, it is not necessarily more significant than print and broadcast media in the formation of national narratives.¹⁸ Public attitudes at times of crisis are also shaped by older mediums, including television news networks, radio shows and even old-fashioned newspapers. Trump’s political rise, for example, owed as much to Fox News as to Twitter. But Twitter, and social media more generally, does have advantages over other forms of political communication in that an individual’s messages can reach a wide audience directly and at great speed. It is easy to access, and accounts are free to create. It makes available multiple and diverse sources of news and analysis to wide audiences and across national boundaries.

Twitter opened for business in 2006, two years after Facebook and a year before Apple launched the iPhone. Its use expanded as smartphones became more common – in 2022 there were some 6.6 billion smartphones in use worldwide, accounting for 80% of all social-media traffic – and according to most estimates there are now more than 370 million Twitter users worldwide.¹⁹ The greatest number of them are in the US (about 65m), followed by Japan (52m), Brazil (17m), the United Kingdom (16m) and India (15m).²⁰ The platform is currently blocked in China, Iran and North Korea.

The world’s third-most visited website, after Google and YouTube, Facebook is even broader in its reach than Twitter, with almost 3bn monthly users in the first quarter of 2023.²¹ In the US it is the most popular social-media platform and a regular source of news for more than 30% of adults.²² Whereas open social-media platforms such as Twitter provide users with the opportunity to interact with strangers, closed platforms such as Facebook, VKontakte and WhatsApp allow users to communicate with people they already know and trust.²³

After initially presenting their platforms as facilitating almost any content that users wished to post, social-media companies and administrators later sought to respond to concerns that this equated to encouraging dangerous and often fake information. Facebook introduced fact-checking policies with the goal of ensuring that shared material was correct and accurate and did not mislead other users.²⁴ Twitter moved in

2020 to remove fake accounts and bots and to prevent inflammatory tweets, especially those that might damage public health by spreading false information about the COVID-19 pandemic. Following Musk’s purchase of Twitter in October 2022, some of these safeguards were removed, and Trump’s account was reinstated. Musk’s arrival as CEO underscored existing questions about the role of private individuals with the ability to manipulate and control the digital information space, working for profit or personal gain.

A similar issue was raised in October 2021 when the ‘Facebook Papers’ were released, including revelations by whistle-blower Frances Haugen. A particularly striking revelation was that although Facebook CEO Mark Zuckerberg had told the US Congress in 2020 that Facebook was managing to remove 94% of hate speech on the platform, internal documents showed the true figure was less than 5%.²⁵ Hate speech noticeably proliferated in the lead-up to elections and during crises, such as the 2019 India–Pakistan crisis. The papers revealed that one of Facebook’s troubling features is the ‘filter bubble’, whereby algorithms consistently feed users more of the same information, creating potentially dangerous echo chambers of misinformation and disinformation.²⁶ The company also prioritised ‘clicks’ over content and would either fail to censor popular individuals and messages or would censor content at the request of a popular individual. As a result, the company and its standards mechanism, XCheck, shielded high-profile individuals, many of whom then used Facebook to spread disinformation or to harass others.²⁷ Despite numerous internal efforts to suppress incendiary posts, the company baulked at content management because of its commitment to neutrality and because it might impede growth.²⁸

In 2021, members of Myanmar’s Rohingya Muslim minority sued Facebook for US$150bn over its alleged role in facilitating genocide, which supposedly included its algorithms amplifying hate speech. Facebook acknowledged in 2018 that it had been negligent in preventing incitement against the Rohingya, noting that ‘Facebook has become a means for those seeking to spread hate and cause harm, and posts have been linked to offline violence’.²⁹ However, genocide has never been dependent on innovative means of communication. The Rwanda genocide of 1994, for example, which was exacerbated by eliminationist rhetoric broadcast by radio, suggests caution should be applied to any judgement that social media plays a unique role in inciting murderous behaviour.

Crisis management in the digital environment

These developments in communications and technology coincided with geopolitical shifts in which China and Russia emerged as strategic adversaries of the Western powers. Strategic reviews by the UK and the US pointed to the need to manage complex multi-domain crises, in which information operations were expected to be an important component. The UK’s 2021 Integrated Review, for example, stated that

technology will create new vulnerabilities to hostile activity and attack in domains such as cyberspace and space, notably including the spread of disinformation online. It will undermine social cohesion, community and national identity as individuals spend more time in a virtual world and as automation reshapes the labour market.³⁰

In the same year, the analyst Margaret Marangione warned of Russia’s and China’s ability to use information operations to exploit groups and individuals, and called for a strengthening of ‘cognitive security’.³¹ The effect of information campaigns may be more acute during international crises, when tensions are high, information might be scarce and time is short. A 2021 report by the Center for Strategic and International Studies, for example, described ‘tactics such as hack and leak operations, forgeries, elite or media co-optation, including flash mobs, bribery, coercion and intimidation, flooding the information zone, false flag operations, causing chaos to provide cover for riskier influence operations, and microtargeting’, and argued that ‘in periods of crisis or conflict, the escalatory potential of such activities may be higher while patterns and pathways of escalation involving influence operations may evade the stepbased and comparatively linear expectations for escalation and crisis management so prevalent among national security decisionmakers’.³² The 2023 ‘refresh’ of the UK’s Integrated Review reported that the Department for Science, Innovation and Technology’s Counter Disinformation Unit ‘will continue to work with both social media platforms and our allies to improve our understanding of the different techniques used in malicious information operations and our ability to counter them, including through the use of intelligence declassification’.³³

To examine the potential escalatory (or de-escalatory) effect of digital information campaigns on international crises, we first develop a framework for thinking about their interaction. This considers the various types of narrative deployed online during crises and how they might affect crisis-management options. We then examine the role of information campaigns in four recent crises: India–Pakistan, 2019; US–Iran, 2020; China’s use of social media during the COVID-19 pandemic, 2020–22; and the conflict in Ukraine from 2013 until the full-blown war following Russia’s invasion in 2022. We intentionally selected cases involving at least one nuclear actor, and crises that had potential to escalate or significantly shift the geopolitical landscape (although obviously the COVID-19 case lacked a military dimension and the death toll was not due to deliberate violence). We conclude by examining trends across the case studies and look ahead to how these might affect future crises. In so doing, we raise concerns about certain trends, including authoritarian actors that rely on social media to spread anti-democratic narratives, using increasingly sophisticated mechanisms; the potential for domestic audiences to become radicalised; and the possibility of information designed for one audience being misinterpreted by another.

Five findings emerge from our analysis of the case studies:

1. Information campaigns, including those on social media, are only one part of wider political struggles shaped by various influential instruments of power. A tweet alone cannot escalate a crisis. Rather, at times of crisis, information campaigns interact with other, more traditional forms of military and economic power.

2. The ease with which public perceptions can be shifted by information campaigns should not be exaggerated, especially when they come from foreign or anonymous sources and challenge established narratives. Crisis messages are likely to be most effective with domestic audiences, where they are likely to reinforce existing views.

3. States cannot therefore be confident that information campaigns will have the desired effects. While conventional wisdom suggests that states such as Russia and China have taken full advantage of the digital environment to advance their information campaigns – and we do not doubt the extent of this effort – there is little hard evidence that they have made much material difference. To the extent that they have done so, it was because they aggravated existing problems in the target society (as with the interference in the 2016 US presidential election the most commonly cited example of success).

4. Information campaigns are limited as strategic instruments because of the difficulty of anticipating the effects that messages will have, even when sent by a political leader. Messages can be interpreted in different ways, and audiences also vary.

5. Social-media campaigns are also hard to control. This is the case even with domestic audiences. They can, for example, create a ‘rally round the flag’ effect that could have unintended international consequences by increasing domestic pressure on an authoritarian leader to escalate.

Two of our findings require additional emphasis. Firstly, many state-led digital information campaigns, such as those of China and Russia, are largely focused on reinforcing narratives among their domestic audiences and have little impact abroad – with important exceptions such as Russia’s interference in the 2016 US election, which successfully reinforced existing tensions in the target country. Secondly, the use of digital information campaigns in the cases we examined were not ‘like war’, as some analysts have characterised the phenomenon, but rather were part of the broader geopolitical picture involving economic and political factors that ultimately determined campaign success or failure.³⁴ The novel technology of social-media platforms might change the way that leaders communicate during a crisis, but the impact of digital information campaigns needs to be considered in context. It remains the case, even in the digital age, that the most important elements determining how crises are managed are the political stakes, along with material factors including military options and economic risks. We do not argue that information campaigns, including the introduction of fake news and malicious rumours, are irrelevant and can have no harmful effects. There is always a possibility that a particular message might have a disproportionate impact, which is why this is an area that needs watching with care. But these effects may be less severe than often feared because, in recent years, there has been an increasing awareness of the pernicious form these campaigns can take and the importance of counteracting them.

Notes

1 Elisa Shearer, ‘More Than Eight-in-ten Americans Get News from Digital Devices’, Pew Research Center, 12 January 2021, https://www.pewresearch.org/fact-tank/2021/01/12/more-than-eight-in-ten-americans-get-news-from-digital-devices/.

2 Marshall McLuhan, Understanding Media: The Extensions of Man (New York: McGraw-Hill, 1964).

3 Purva Grover, Arpan Kumar Kar and Yogesh Dwivedi, ‘The Evolution of Social Media Influence – A Literature Review and Research Agenda’, International Journal of Information Management Data Insights, vol. 2, no. 2, November 2022, https://doi.org/10.1016/j.jjimei.2022.100116.

4 Constance Duncombe, ‘The Politics of Twitter: Emotions and the Power of Social Media’, International Political Sociology, vol. 13, no. 4, 13 August 2019, pp. 409–29, https://doi.org/10.1093/ips/olz013.

5 See, for example, Peter W. Singer and Emerson T. Brooking, ‘What Clausewitz Can Teach Us About War on Social Media’, Foreign Affairs, 4 October 2018, https://www.foreignaffairs.com/articles/2018-10-04/what-clausewitz-can-teach-us-about-war-social-media.

6 Twitter, ‘Permanent Suspension of @realDonaldTrump’, 8 January 2021, https://blog.twitter.com/en_us/topics/company/2020/suspension.

7 This figure includes retweets. Tweets sent during Trump’s presidency are available in the Trump Twitter Archive: https://www.thetrumparchive.com. He averaged 5.7 tweets per day during the first six months in office but this increased to 34.8 per day during the second half of 2020. Niall McCarthy, ‘Chart: While He Still Could, How Much Did Donald Trump Tweet?’, The Wire, 12 January 2021, https://thewire.in/world/donald-trump-twitter-chart.

8 Donald Trump (@realDonadTrump), tweet, 2 January 2017, https://www.thetrumparchive.com/.

9 Donald Trump (@realDonalTrump), tweet, 3 July 2017. The subsequent tweet reads ‘ … and Japan will put up with this much longer. Perhaps China will put a heavy move on North Korea and end this nonsense once and for all!’, 3 July 2017, https://www.thetrumparchive.com/.

10 See, for example, Peter Baker and Choe Sang-Hun, ‘Trump Threatens “Fire and Fury” Against North Korea if It Endangers U.S.’, New York Times, 8 August 2017, https://www.nytimes.com/2017/08/08/world/asia/north-korea-un-sanctions-nuclear-missile-united-nations.html.

11 See, for example, Anna Fifield, ‘Kim Jong Un Calls Trump a “Mentally Deranged Dotard”’, Washington Post, 21 September 2017, https://www.washingtonpost.com/news/worldviews/wp/2017/09/21/north-korean-leader-to-trump-i-will-surely-and-definitely-tame-the-mentally-deranged-u-s-dotard-with-fire/.

12 Bob Woodward, Fear: Trump in the White House (New York: Simon and Schuster, 2018), p. 300.

13 Trump was probably referring to numerous recent failed missile tests by North Korea (18 out of 86). Donald Trump (@realDonaldTrump), tweet, 2 January 2018, https://www.thetrumparchive.com.

14 Woodward, Fear: Trump in the White House, p. 302.

15 Jeffrey Lewis, The 2020 Commission Report on the North Korean Nuclear Attacks Against the United States: A Speculative Novel (Boston, MA: Mariner Books, 2018).

16 Chrystia Freeland (@cafreeland), tweet, 2 August 2018, https://twitter.com/cafreeland/status/1025030172624515072?lang=en.

17 Andrew England and Simon Kerr, ‘Saudi Arabia’s Furious Attack on Canada Shocks Western Allies’, Financial Times, 7 August 2018, https://www.ft.com/content/688805d0-9a55-11e8-9702-5946bae86e6d?.

18 Amy Mitchell, Elisa Shearer and Galen Stocking, ‘News on Twitter: Consumed by Most Users and Trusted by Many’, Pew Research Center’s Journalism Project, 15 November 2022, https://www.pewresearch.org/journalism/2021/11/15/news-on-twitter-consumed-by-most-users-and-trusted-by-many.

19 ‘Number of smartphone mobile network subscriptions worldwide from 2016 to 2022, with forecasts from 2023 to 2028’, Statista, 30 March 2023, https://www.statista.com/statistics/330695/number-of-smartphone-users-worldwide; and ‘Twitter Users, Stats, Data, and Trends’, Datareportal, updated 11 May 2023, https://datareportal.com/essential-twitter-stats.

20 Ibid.

21 ‘Number of monthly active Facebook users worldwide as of 1st quarter 2023’, Statista, 9 May 2023, https://www.statista.com/statistics/264810/number-of-monthly-active-facebook-users-worldwide.

22 ‘Social Media and News Fact Sheet’, Pew Research Center, 20 September 2022, https://www.pewresearch.org/journalism/fact-sheet/social-media-and-news-fact-sheet.

23 Vipin Narang and Heather Williams, ‘Thermonuclear Twitter?’, in Vipin Narang and Scott Sagan (eds), The Fragile Balance of Terror (Ithaca, NY: Cornell University Press, 2023), pp. 63–89.

24 ‘Understanding Facebook’s Fact-checking Program’, Meta, November 2019, https://www.facebook.com/gpa/blog/misinformation-resources#:~:text=Facebook%20started%20its%20fact%2Dchecking,40%20languages%20around%20the%20world.

25 Cristiano Lima, ‘A Whistleblower’s Power: Key Takeaways from the Facebook Papers’, Washington Post, 26 October 2021, https://www.washingtonpost.com/technology/2021/10/25/what-are-the-facebook-papers/.

26 Sally Adee, ‘How Can Facebook and Its Users Burst the “Filter Bubble”?’, New Scientist, 18 November 2016, https://www.newscientist.com/article/2113246-how-can-facebook-and-its-users-burst-the-filter-bubble/.

27 Jeff Horwitz, ‘Facebook Says Its Rules Apply to All. Company Documents Reveal a Secret Elite That’s Exempt’, Wall Street Journal, 13 September 2021, https://www.wsj.com/articles/facebook-files-xcheck-zuckerberg-elite-rules-11631541353?mod=article_inline. See also Jeff Horwitz, ‘The Facebook Files: A Wall Street Journal Investigation’, Wall Street Journal, https://www.wsj.com/articles/the-facebook-files-11631713039.

28 Ibid.

29 Dan Milmo, ‘Rohingya Sue Facebook for £150bn over Myanmar Genocide’, Guardian, 6 December 2021, https://www.theguardian.com/technology/2021/dec/06/rohingya-sue-facebook-myanmar-genocide-us-uk-legal-action-social-media-violence.

30 UK Cabinet Office, ‘Global Britain in a Competitive Age: The Integrated Review of Security, Defence, Development and Foreign Policy’, 2 July 2021, https://www.gov.uk/government/publications/global-britain-in-a-competitive-age-the-integrated-review-of-security-defence-development-and-foreign-policy/global-britain-in-a-competitive-age-the-integrated-review-of-security-defence-development-and-foreign-policy.

31 Margaret S. Marangione, ‘Words as Weapons: The 21st Century Information War’, Global Security and Intelligence Studies, vol. 6, no. 1, Spring/Summer 2021, p. 149, https://gsis.scholasticahq.com/article/25512-words-as-weapons-the-21st-century-information-war.

32 Rebecca Hersman et al., ‘Influence and Escalation: Implications of Russian and Chinese Influence Operations for Crisis Management’, Center for Strategic and International Studies, 9 November 2021, https://www.csis.org/analysis/influence-and-escalation-implications-russian-and-chinese-influence-operations-crisis.

33 UK Cabinet Office, ‘Integrated Review Refresh 2023: Responding to a More Contested and Volatile World’, 13 March 2023, https://www.gov.uk/government/publications/integrated-review-refresh-2023-responding-to-a-more-contested-and-volatile-world.

34 See, for example, Peter W. Singer and Emerson T. Brooking, Like War: The Weaponization of Social Media (Boston, MA: Houghton Mifflin Harcourt, 2018).