Written evidence submitted by Dr Stephen Hutchings, Dr Vera Tolz, Dr Alexander Voronovici (DIS0022)

Disinformation diplomacy:

How malign actors are seeking to undermine democracy

This evidence was prepared by members of the 3-year research project, (Mis)Translating Deceit: Disinformation as a Translingual, Discursive Dynamic, funded by UKRI and led from the University of Manchester (PI – Stephen Hutchings; Lead Co-I: Vera Tolz; Researcher: Alexandr Voronovici) in partnership with Loughborough and Leeds Universities and Chatham House.

Focused primarily on Russia, our evidence offers a corrective to certain flaws and blind spots in current counter-disinformation practices. It stresses the need to pay more attention to accurate assessment of the impact of foreign influence operations and their place within broader state strategies. It recommends that more account be taken of differences between Russian strategies aimed at western countries and those targeting Eastern Europe and the Global South, but also of the messy entanglement of foreign with domestic disinformation. It highlights the importance of precision in counter-disinformation analysis and, of recognizing how disinformation and counter-disinformation shape one another. Finally, it calls for more in-depth reflection on the complex relationship between counter-disinformation activities and the democratic systems they strive to defend. Beginning with some general observations, we next summarise earlier research on the impact of Russia Today (RT), still highly pertinent owing to the outlet’s recent re-orientation to non-Western audiences, before turning to Russia’s post-2022 European propaganda strategy, and concluding with comments on counter-disinformation practices and some final suggestions.

Kremlin Influence Operations: General Observations and Background

Kremlin international propaganda and disinformation are at the centre of Western security concerns. Nevertheless, it is important to distinguish threats from malign states like Russia to citizens in the UK from those posed to UK interests in Global South. Moreover, disinformation disseminated to UK audiences by openly Kremlin-affiliated outlets like Russia Today (RT) or Sputnik (pre-2022) and by clandestine Kremlin proxy actors (post-2022) have limited effect since their output has minimal following. In the UK and most EU countries, Russia’s public image is poor and there is a widespread mistrust of its narratives.

However, Russia’s propaganda efforts around the war in Ukraine deserve special mention:

• Some UK and EU citizens disapprove of the level of state support for Ukraine (e.g. a significant proportion of Reform UK voters), reflecting a belief that this support detracts from their own economic wellbeing[1]
• Despite decreases in overall UK and European popular support for Ukraine in the past two years, it remains at high levels
• Opposition to such support results much less from direct Russian influence operations than from the positions of domestic right-wing populist parties (e.g. Reform UK or Alternative fur Deutschland)
• Collaboration, overt and covert, between such parties and Russia-affiliated actors is thus a major concern

The direct threat from Russian propaganda and disinformation varies by region:

• It is limited in the UK and West European states
• It is much higher in countries neighbouring Russia and having large Russian-speaking populations (Estonia and Latvia, Ukraine, Belarus, Moldova; Central Asia; the Caucasus)
• Russian disinformation activities in Africa, Asia, the Middle East and Latin America are particularly challenging due to historically rooted mistrust of the West in general and the UK specifically; Russian campaigns exploit this effectively
• Via non-European and non-Western sources, Russian narratives can find their way indirectly back into Western media spaces; such sources should continue to be monitored

The Russian influence and disinformation activities we analyse should not be conflated with hacking and cyber-attacks. Beyond our remit, these pose a much graver threat to the UK domestically and internationally.

Assessing RT’s Impact in 2018-2020 and 2022

Our analysis of @RT_com’s English Twitter audiences and RT’s Facebook’s English, Spanish and Arabic pages is based on data gathered between 2018 and 2020 and in October 2022.[2]

2020 @RT_Com Twitter analysis

• Account had 2.6 million followers in 2020
• We selected a sample of 10, 391 followers (regarded as representative)
• Engagement was low
• Only 0.15% of engagement tweets related to RT’s content
• 10 users accounted for 40.4 percent of the engagement
• According to Botometer, 39% of @RT_com followers’ accounts from our sample were flagged as bots
• Thus, the findings above and below should all be treated with caution
• Most engagement was around human-interest stories, not political content

2020 RT Facebook pages

• RT Arabic had 16.7 million followers (as of April 2020)
• RT Spanish – 16.4 million
• RT (main Facebook page in English) – 6 million
• Prior to the war in Ukraine in 2022 there was a high level of engagement across all Facebook pages
• RT Spanish, targeting Latin America, performed best on the most meaningful form of engagement – shares
• RT Spanish Facebook page –most engagement was with posts of an anti-American character, or with human interest stories
• Elsewhere most engagement was with human interest news

Comparative context

• RT (English) Facebook page had 6 million followers – the lowest following among Facebook pages of other international broadcasters used for comparison
• Al Jazeera (14 million), CNN (19 million), BBC (53 million), CGTN (106 million)
• Engagement weighted per million followers: Al Jazeera had the highest median engagement, with 92 engagements per million; the BBC was second with 89 per million; RT third – 77 per million; CGTN lowest
• RT Arabic’s Facebook page had the second largest following among international broadcasters (17 million)
• Al Jazeera – 24 million; BBC – 12 million; CNN - 3 million; CGTN - 15 million
• Follower engagement with content was very low across all broadcasters (as typical for social media in the Arab world)
• RT Spanish’s Facebook page was in second place with 16 million followers
• CGTN – 17 million; CNN – 13 million; Al Jazeera – 4 million; BBC – 5 million
• RT Spanish – third in median engagement and meaningful engagement (comments/shares) per million after AJ Spanish and BBC Spanish
• RT Spanish’s Facebook page appears to be the most successful of all RT’s social media platforms – a factor with huge relevance to the post-2022 landscape

October 2022

• Unsurprisingly, RT English, Arabic and Spanish Twitter followings showed no major changes, despite Western bans (RT’s primary audiences had always been non-Western).
• RT’s multilingual Facebook pages also experienced minimal changes, though engagement dropped significantly across the board
• RT Arabic’s Facebook page lost a quarter of its followers in one day in September 2022, following META’s operation targeting non-genuine Russia-related accounts
• However, in October 2022 RT Arabic Facebook’s Ukraine-war related posts enjoyed very high engagement levels (assuming we can trust the figures) 
• RT thus appears to have had success in promoting Kremlin war narratives to its international audiences

Kremlin-Sponsored Proxy Media Outlets post-2022

Following Western access bans on Russia’s primary foreign propaganda tools, new channels of Kremlin influence, particularly proxy media outlets, emerged. These:

• Disseminate propaganda in multiple languages
• Mimic local (UK, European, North American) media providers
• Promote the content of banned Russian entities, enabling pro-Kremlin content to penetrate Western media spaces
• Attracted much Western attention owing to their deceitful character which differentiates them from more traditional outlets with open connections to the Kremlin

We analysed five confirmed Kremlin-sponsored proxy outlets – Reliable Recent News (RRN), Berliner Tageszeitung (BTZ), Voice of Europe (VoE), France et EU, and a network of multilingual Pravda websites.[3]

Our study found:

• A disconnect between the substantial resources Russia commits to these operations and the minimal level of influence and traction they achieve; according to tools analysing references to proxy websites via hyperlinks, VoE had only ca1000 referring domains in summer 2024, while RRN even less – 185 referring domains
• Audience engagement and penetration of western media spaces remain modest
• Even the most prominent Russia-sponsored proxy outlets cannot compare with RT impact-wise, even taking account of Western sanctions on RT since 2022 (VoE’s Twitter following – the highest among the proxies we studies stood at ca180 thousand; compared to RT’s 3.2 million)
• RT’s and Sputnik’s profile and reach were – and are – still much greater than those of any of the propaganda tools which have replaced them in European media spaces

Our findings suggest the Kremlin may be shifting towards a ‘swamp and distract’ strategy, prioritising the generation of large volumes of content over impact, aiming to overwhelm the information space, making it difficult to distinguish legitimate news from propaganda, and triggering the dedication of considerable Western resources to investigate them.

Actors Spreading Disinformation and Propaganda

The coordinated restrictions imposed on RT and Sputnik have a patchy record of success:

• RT content is still available in the UK via old versions of its smartphone app
• Certain Russian state news websites remain accessible, including TASS (English), and that of Channel 1, whose news bulletins can be watched freely on iPhones using Safari
• Some RT spin-offs based on individual programmes maintain online presences in Western countries, evading restrictions applied to the RT “mothership”
• RT and Sputnik have, since 2022, accelerated their pivot to the Global South – long seen as a more promising environment for counter-hegemonic narratives – setting up a new African operation, for example

Besides explicitly state-affiliated actors, we analysed a complex web of actors with different levels of affinity to Moscow and part of a sprawling Russian state apparatus which extends into a grey zone that includes:

• corrupt businesspeople
• individual non-Russian Kremlin sympathisers
• former intelligence operatives
• semi-criminal entrepreneurs with largely financial incentives.

An important dimension of this mosaic are groups and individuals from Western counties including:

• populist right-wing parties (Reform UK, Alternative fur Deutschland) with whom Russia-affiliated actors collaborate (e.g. in Germany during elections)
• disgruntled former officials from government departments (e.g. Craig Murray, UK’s former ambassador to Uzbekistan; Paul Craig Roberts, who held a sub-cabinet office in the US federal government and is now a leading conspiracy theorist)

At the same time, using qualitative methods we compared the output of our selected Kremlin proxies with ‘domestic’ disinformation spreaders in Europe and North America lacking proven Kremlin links. While there was a large diversity in approach across the Kremlin- sponsored outlets, overall, the domestic actors strayed far further from the established facts. What is routinely described as Russian ‘disinformation’ often falls well short of the term’s full definition and is better characterised as ‘hyper-partisan’ reporting.

Channels and Tools of Kremlin’s Proxy Media Outlets

There are additional tools and technologies that both facilitate Kremlin proxy activities and became instrumental for other channels of disinformation.

Among these are:

• the growing availability of AI technologies capable of generating large volumes of aggregated news content to boost website profiles and authenticity
• auto-translation tools that can rapidly render vast amounts of shared content in multiple languages; the Pravda multilingual network is a good example, but it is poorly coordinated and does not seek to disguise the Russian origins of its material
• the rise of Telegram instant messaging/social networking channels as a dissemination and amplification platform relatively free of regulatory oversight
• Musk’s virtual abandonment of Twitter’s hate speech and false content rules (with posts by RT affiliates, including RT India, now appearing in X feeds)
• ‘Troll factories’, fake accounts, and bots operating on social media and website comment sections that promote specific propaganda and disinformation narratives, creating a sense of ‘popular’ support for Kremlin-friendly positions and disapproval of unfavourable ones

However, Russian state communication strategists do not currently seem able to exploit AI and auto-translation tools to their full potential. The quality of the auto-translations is  mixed, some translations being riddled with linguistic errors. Nor has Telegram yet proved a viable substitute for Twitter, Facebook and Instagram whose use in Western media spaces by RT and Sputnik proved relatively successful (if not at the levels feared by many democratic commentators). RT’s Telegram channel has ca200 thousand subscribers compared with more than 3 million on Twitter.

Kremlin-sponsored outlets quote Western mainstream media to authenticate their own agenda, and to gain traction with followers of reputable sources. In some cases, mainstream Western media become unwitting sources for Kremlin disinformation. For instance, reports from Politico or Euronews were copied almost verbatim with minor linguistic tweaks or new closing sentences that align then with the proxy outlet’s agenda. We identified such examples for news about Ursula von der Leyen's rivalry with Charles Michel,[4] and the financial consequences of Macron calling snap elections.[5] Such examples are rare, however.

Counter-Disinformation Efforts

While some counter-disinformation efforts, particularly against Russian disinformation were successful, overall, the record is mixed. The problems result from the uncontrolled speed with which the counter-disinformation industry has grown. Falling into three categories they can be rectified if governments conduct a coordinated, cross-institutional audit of counter-disinformation activities leading to new approaches informed by thought-through, systematic modes of analysis.

Problems reflecting counter-disinformation’s conflicted relations with democracy

• The counter-disinformation landscape is shaped by its key actors’ portrayal of their relationship to democracy: Counter-disinformation units (CDUs) highlight their independence as a democratic virtue, yet scrutiny of their funding arrangements regularly reveals significant government/state support, conflicting with their rhetoric of transparency and independence and creating the impression that they are state proxies.
• The genuine, and constant, funding needs of many CDUs encourages inflation of the disinformation threat they monitor. The EU’s East Stratcom Taskforce (EST), for example, regularly highlights the volume of disinformation instances it identifies, including within its statistics instances that fail to meet its own definitions
• Conversely, CDUs run directly by governments risk fostering public perceptions that they are sinister state surveillance operations, as the UK government learned when it mistakenly targeted a mainstream politician for spreading disinformation (https://www.theguardian.com/politics/2023/jun/18/caroline-lucas-flagged-by-disinformation-unit-over-covid-criticism)
• The emphasis many CDUs place on the democratizing role of bottom-up, civic participation in their work is in tension with their need for professional, impartial expertise. EU Disinfo Lab celebrates its membership of ‘a vast, passionate community’, whilst stressing its commitment to dispassionate analysis
• CDUs are vulnerable to shifts in the democratic consensus; the emphasis placed by GDI and the US-based Digital Forensic Research Lab (DFRL) on minority rights, diversity and inclusion and on rejecting hate speech now clashes with mainstream objections to ‘cancel culture’ and ‘free speech’ suppression (cf. Trump’s assaults on disinformation analysts)

Problems resulting from the origins of counter-disinformation

• The term ‘disinformation’ originates in political polemic, not professional, impartial knowledge production. A negative consequence of its continued centrality to CDU analysis is its misappropriation by non-democratic rivals both within democracies (the populist right) and beyond them (authoritarian state versions of CDUs)
• The key historical rivalry from which disinformation lexicon developed was that of the Cold War whose binary structures make for blunt analytical tools, resulting in the construction of ‘disinformation’ as the single antithesis to disparate liberal democratic values which do not always align (commitment to minority rights; faith in professional expertise; the sanctity of judicial processes)
• The lingering Cold War legacy of the current CDU analytical approach explains the skewed emphasis on ‘foreign interference’ (cf. the flawed FIMI apparatus) which overlooks the domestic dimension to disinformation spreading and can misdirect attention toward states when complex networks of substate and non-state networks are at play (e.g. confusion over identification of the source of the disinformation that prompted the Southport riots)

Problems with the conceptual approach favoured by CDUs

• CDUs are attached to unwieldy taxonomies listing multiple ‘techniques’ and ‘tools’ (An EST report enumerates as many as 250). As well as displaying incompatibility (‘creating a fake account’ is different from ‘amplifying a conspiracy narrative’, for example), these tools belong to linear models assuming agent-technique-recipient processes which do not correspond to how false narratives emerge and circulate
• Lingering cold war binarism is responsible for the continued influence of crude false/true approaches to information manipulation which in fact requires a more layered understanding. This is linked to imprecision in determining the nature and source of the deception (the use of loose hybrid terms like ‘disinformation-propaganda’; the identification as state disinformation of Kremlin-sponsored news reports which quote rather than endorse discredited accounts)
• Growing CDU awareness of these difficulties has prompted a shift away from tracking instances of ‘deliberate falsification’ (relatively rare and difficult to prove) to the normative rejection of objectionable, ‘adversarial narratives’ (now the favoured concept of the GDI). However, this move ignores the fact that most politics, democratic and anti-democratic, is defined by the clash of adversarial narratives and the creation of ‘in’ and ‘out’ groups

Suggestions

• The low direct impact of Russia’s new networks of proxy outlets in the West suggests that policymakers should balance addressing the threats these pose with the danger of overinflating their risk, which may inadvertently fulfil Kremlin goals. Overstating their significance can give them more publicity; distracting democratic states from true threats to security; and corroborating Russian state narratives regarding Western “Russophobia”, liberal democratic “cancel culture” and “suppression of free speech”. It also contributes to the impression, favourable to Kremlin goals, that no media can be trusted
• Successful bans applied in Western media spaces do not prevent Kremlin narratives from re-entering those spaces indirectly via Russia’s large non-Western audiences whose role in facilitating Kremlin goals must be studied carefully; RT and Sputnik remain major players in this regard, despite their relative inaccessibility in the West
• What is required of counter-disinformation units is the development of a fresh apparatus grounded in a sophisticated conceptualization of the complex relationship between truth and politics; a capacity to distinguish reality from fact, fact from knowledge, and knowledge from societal conviction; and attention to basic linguistic differences like that between claim and reported claim
• Such distinctions cannot always be established through the quantitative methods and computer-driven tools that continue to dominate disinformation analysis. Our research suggests that caution should be exercised when applying the latest tools of detection and analysis to digital propaganda. Automated tracking software can give highly misleading results, if not accompanied by qualitative, manual analysis

24 February 2025