IN THIS EDITION
ON THE RISE
Weekly Watch of Emerging Disinformation Risks
AI-Political Influencers: the New Gods of Propaganda and Disinformation?
A crucial check on fanaticism has always been the simple fact that its prophets rarely manage to live up to their own public sermons. It is far easier to preach the equality of mankind and the virtues of socialist austerity than to practice either, once power has been secured. Easier, too, to champion family values, religion and traditional morality from a podium than to genuinely renounce every material temptation in private life.
Artificial Intelligence, however, may have handed extremists a new weapon, one capable of making absolutist propaganda more effective, or at least less vulnerable to collapsing under the weight of ordinary human hypocrisy: synthetic influencers.
One of the more revealing cases comes from Britain, where the far-right, anti-immigration rapper Danny Bones has been amassing followers online through a stream of racist and incendiary content. He was among the “stars” featured at the far-right Unite the Kingdom rally held in London on May 16. The catch is that Danny Bones does not exist. He is an AI-generated character created by The Node Project and funded by the extremist group Advance UK. Some followers are unaware of the deception (the quality of the images is, for now, far from perfect), while many others simply do not care. If the message resonates, the logic goes, why should it matter whether the messenger is artificial?
There are several answers to that question, but one stands out. In matters of ethics and morality, public figures should not demand sacrifices from others that they are unwilling (or unable) to make themselves. Since genuine Savonarolas are rare, ideological movements have often collapsed precisely where personal conduct failed to match public purity. AI offers a way around that problem.
Danny Bones – and his future epigones – will never be caught taking bribes. He will never stumble drunkenly out of a nightclub in compromising company. He will never make an off-colour remark on live television. He will never fall in love with a Muslim woman. He will never mellow with age or drift away from the doctrine he promotes. Danny Bones is, and will remain, perfectly aligned with his message.
More importantly, thanks to AI he can possess the gifts of ubiquity and near-omniscience. He can be everywhere at once, comment on every issue, appear at a hundred events simultaneously, all without the slightest deviation from ideological orthodoxy. He can hold thousands of personalised conversations at the same time, telling each listener exactly what they want to hear, seducing, persuading and radicalising at industrial scale. And Danny Bones is merely an early example. The phenomenon is still in its infancy, at least in terms of the political potential AI may unlock in the years ahead.
The technology does not benefit all political actors equally. Not every politician seeks to impose some sweeping code of moral purity on society. But in the age of social media and algorithmic amplification, the advantage for extremists is enormous. Outrage, verbal violence and polarization generate clicks; algorithms reward them; platforms and influencers profit from them. Hatred and division spread with the efficiency of a contagion. And soon, those who manufacture and weaponise such rhetoric may no longer need to worry about living consistently with it. AI can do the job for them: manufacturing flawless two-dimensional heroes, aesthetically polished and ideologically immaculate, as persuasive in their synthetic perfection as the printed propaganda icons of the twentieth century’s totalitarian regimes.
ZOOM-IN
A Closer Look at Cases Detected by the EDMO Network
In Jack Finney’s classic 1954 sci-fi horror novel, The Body Snatchers, alien seeds from space grow into plant-like pods that secretly replicate and replace human beings with emotionless clones while they sleep. Something similar seems to happen today with the widespread adoption of generative AI in political communications. Synthetic political influencers – flawless and infallible – duplicate human knowledge, memories, aspirations and needs, garnering attention and peddling propaganda on social media. Several investigations show the harmful dimensions of the phenomenon, as well as signs of societal frustration with both the technology and the unwillingness of the tech industry to mitigate the risks it generates.
How Synthetic Content Transforms Political Communications
The use of synthetic content for political communication has steadily increased during recent months. EDMO and Facta examined this phenomenon one year ago, in the aftermath of the German federal elections, and unveiled how far-right political movements across Europe were increasingly using AI-generated images and videos to influence voters. They reported about campaigns in Germany, France, Poland, Italy, Ireland, and Spain, where politicians and activists used synthetic visuals to idealize a nostalgic vision of “traditional” Europe. Examples included Germany’s AfD sharing AI-created scenes contrasting blond families with dangerous migrants, and Italian politicians depicting dystopian futures dominated by Islamic imagery. As stressed in the article, AI tools are especially effective for emotional, visually engaging propaganda on social media, where memes and trends help extremist messages spread widely. These materials are often unlabeled, making them misleading and potentially violating the Election Integrity Guidelines under the Digital Services Act.
Since last year, AI-powered political propaganda has grown in volume, intensity and sophistication, as revealed by recent investigations by GADMO – concerning a large network of AI-generated female personas on Instagram, TikTok, and Facebook promoting the AfD party – and by NORDIS, which unveiled an AI-generated “influencers” campaign for Trump ahead of US midterms.
Fear, Boos and Firebombs: AI Creates Growing Anger Among Americans
The article examines growing public dissatisfaction in the United States with the rapid expansion of artificial intelligence. While AI is promoted as a transformative technology, many Americans increasingly associate it with job insecurity, rising energy consumption, and pressure on natural resources such as water, raising environmental concerns in communities hosting this infrastructure, and sparking fears among workers that automation and generative AI tools could replace jobs across sectors, from administration to creative industries.
At the same time, constant exposure to inflammatory AI-generated material contributes to growing hostility and anxiety among Americans by making public discourse more aggressive and by reducing trust in media, institutions, and fellow citizens. Examples include fake scenes of violence, disasters, and political unrest that circulate rapidly on social media before fact-checkers can respond. These anxieties are contributing to broader frustration with the tech industry and skepticism toward political leaders who strongly support AI development.
The article suggests that AI may become an important political issue in upcoming U.S. midterm elections, as debates over employment, energy use, environmental costs, and regulation increasingly shape public opinion and voter concerns.
Meta Oversight Board is Investigating AI Video Impersonating Péter Magyar
The article reports that Péter Magyar became the subject of a review by the Meta Oversight Board over an AI-generated Facebook video impersonating him during Hungary’s election period. The video, posted by pro-government influencer István Szakács, showed a synthetic version of Magyar angrily reacting to a political controversy connected to leaked personal data. More than 200 users reported the clip, but Meta decided neither to remove it nor label it as AI-generated, arguing that it was satirical and unlikely to mislead the public.
The case has become part of a broader debate about how social media platforms should handle AI-generated political content, especially during elections. The Oversight Board launched a public consultation on the risks posed by deepfakes, political influencers, and coordinated disinformation campaigns. Critics argue that realistic AI impersonations can manipulate voter perceptions even when presented as humor, raising concerns about election integrity, platform accountability, and the effectiveness of current moderation policies.
Such concerns are echoed by an article from Kallkritik Byran, which examines how altered election posters, fabricated statements and fake social media posts falsely attributed to politicians are increasingly being used to distort political messaging in Sweden and elsewhere in Europe. It describes how manipulated campaign material often imitates official political communication, making it difficult for voters to distinguish satire, activism, and deliberate disinformation.
ELECTION BEAT
Tracking electoral disinformation through EDMO Hubs
Malta General Election 2026 – Weekly Brief #3: Unearthing a Pandora’s Box
In the final week before Malta’s general election, disinformation, AI-generated content, misleading political claims, and suspicious online activity are increasingly shaping the campaign environment. The third weekly brief by MedDMO examines key developments affecting Malta’s information space.
One controversy emerged after Prime Minister Robert Abela claimed during a debate that Malta already provides all the best cancer treatment medication through the public healthcare system. The brief argues this is misleading, noting that some costly treatments are not covered by the state.
The report also highlights concerns surrounding prediction markets on Polymarket, where users can bet on election outcomes despite election betting being illegal in Malta. Experts warn such platforms could influence public perceptions of the race.
Artificial intelligence is also becoming more visible in campaigning. The Partit Nazzjonalista launched an AI chatbot to explain its manifesto, but it generated inaccurate visuals and misleading content before its image-generation function was disabled.
The brief also raises concerns about suspicious websites and coordinated “troll farming” activity amplifying pro-government narratives.
Sandygate: The “True” Crime Case That Shook Up the Election Campaign and Public Debate in Cyprus
The “Sandygate” case in Cyprus emerged during the run-up to the 24 May 2026 parliamentary elections and quickly evolved from a political scandal into a broader dispute over the reliability of digital evidence. The allegations were brought forward by journalist and Volt Cyprus candidate Makarios Drousiotis, who accused senior political, judicial and economic figures of involvement in corruption and an influence network centred on a woman referred to as “Sandy.” Initially, the case gained traction through allegations of abuse and claims of a wider “Brotherhood” network, supported by messages, audio recordings, and images.
However, Cypriot media scrutiny soon shifted attention from the accusations themselves to the authenticity of the supporting material. Fact-checking investigations by outlets such as AlphaNews identified multiple inconsistencies, including images traced back to unrelated websites, audio clips likely sourced from documentaries, and questionable text message formatting. The case had direct political consequences, influencing party positioning and intensifying debate around corruption, institutional trust and media credibility. While the truthfulness of the allegations remains unresolved, Sandygate highlights how disputed digital evidence and information manipulation can rapidly shape public debate in Cyprus’ already polarised online environment, which has also faced concerns over foreign influence, including activity linked by Meta to Chinese state-controlled media.
GLOBAL PULSE
Disinformation narratives shaping the world’s conversations
Flotilla Goals Misrepresented: Fact Checks Challenge Meloni’s Claims and Highlight Legal Dispute
The IDMO fact‑check argues that Giorgia Meloni’s claim about the Global Sumud Flotilla is misleading: she suggested the mission falsely presented itself as humanitarian while actually aiming to challenge Israel’s naval blockade. In reality, the organisers had been transparent from the start that their goals were both to deliver aid to Gaza and to politically contest the blockade, meaning there was no hidden agenda – only a dual humanitarian and protest purpose publicly stated in advance.
A related analysis by DE FACTO provides legal context, noting that the flotilla indeed aimed to bring humanitarian supplies and break what activists consider an unlawful blockade, reinforcing that these objectives were explicit. They add that Israel’s interception of the vessels – carried out in international waters – raises serious legal concerns under the law of the sea, highlighting how the controversy centres not just on political claims but also on competing interpretations of international law.
Misleading Ebola Claims Spread as Conflict and Misinformation Fuel Deadly Outbreak in DR Congo
The Newtral fact‑check explains that viral claims about “Ebola reaching the United States” are misleading: the case in question involves a US doctor who was infected while working in the Democratic Republic of Congo, not a domestic outbreak. He was evacuated to Berlin for treatment and remains in stable condition, while several close contacts were also transferred to Europe as a precaution. Health authorities stress that the broader situation is a regional outbreak (declared an international emergency, but not a pandemic) and that the risk outside affected areas remains low.
Reporting from CNN and other outlets situates this case within a wider crisis: a rapidly growing Ebola outbreak in eastern DRC, driven by the rare Bundibugyo strain, with hundreds of suspected cases and spread to neighbouring countries. Containment is complicated by conflict, displacement and weak health infrastructure, making it difficult to trace contacts and isolate patients. At the same time, a Balobaki Check analysis highlights how misinformation worsens the situation on the ground, with some communities attributing the disease to curses or denying it exists, leading to distrust of medical teams and faster spread of the virus.
False Claims About Ukraine Drone Attacks From Latvia Debunked as Russian Disinformation
The article debunks claims circulating on social media and promoted by Russian officials that Ukraine is preparing to launch drone attacks against Russia from Latvian territory. The claim originated from statements at the UN by Russia’s representative and from Russia’s foreign intelligence service, but the article finds no evidence to support it and classifies it as misleading and taken out of context.
Officials from Latvia, NATO, and Ukraine all strongly deny the allegation, stressing that Latvia would not allow its territory to be used for attacks against another country. The Latvian president explicitly called the claim false, while Ukrainian officials described it as part of a broader Russian propaganda effort aimed at destabilising public opinion in the Baltic region. The article concludes that the narrative is baseless and part of recurring disinformation portraying NATO as an aggressor.
ON A DIFFERENT NOTE
AI influences democracy through multiple entry points, including elections, citizen deliberation, government services, and social cohesion, all of which are influenced by geopolitics and security. All of these domains, mapped in this paper, face threats related to influence, integrity, and bias, yet also present opportunities for targeted interventions.
Carnegie report AI and Democracy: Mapping the Intersections
Paolo Cesarini, Editorial Director
Tommaso Canetta, Editor-in Chief
Editorial Staff: Elena Coden, Paula Gori, Elena Maggi
This edition draws in part on automated translation and reflects information available as of 28 May 2026. Later developments may not be included.
