The social media platform X has for months been aflush with praise for the United Arab Emirates and Azerbaijan in posts shared by hundreds of profiles – and all found to be fake.
Analysed by AFP over several days, the large-scale operation, powered by artificial intelligence, points to a sophisticated, coordinated influence campaign not unlike those carried out by Russia in recent years – though its instigator and objectives remain unclear.
“French people, (if you are) disappointed and weary of endless debate, it is time to think about an alternative, such as the United Arab Emirates, where business conditions are really attractive,” one sham account wrote.
It is part of a network of more than 2,300 active accounts in a dozen languages, including English, French and German, according to the collective Antibot4Navalny which monitors influence operations on X and teamed up with SourcesOuvertes to identify the campaign. Some boast up to a few hundred followers.
To accrue visibility and credibility, the bogus profiles comment on posts by mainstream media, local news organisations, and influential accounts.
While they discuss a breadth of topics, what they have in common is their fervent acclaim for the UAE and Azerbaijan, and particularly their flourishing economies.
The profiles also agree on a “desire to change elected officials, established institutions and status quo” in the West, with the Emirati model depicted as “the most successful alternative”, said Antibot4Navalny in a series of posts published on X last week.
Some convey political messages, such as “Nagorno-Karabakh is part of Azerbaijan”, and Baku “has the right to claim its occupied territories”, referring to the sovereignty conflict over the disputed region, while others support Azeri athletes.
Former Soviet republics Azerbaijan and Armenia have seen decades of war and tension over the breakaway ethnic Armenian province, recognised as part of Azerbaijan.
The international community has been ramping up pressure for an agreement between the neighbours before the COP29 summit later this month.
COP hosts
The disinformation efforts come as the UAE and Azerbaijan were chosen to host the COP28 climate change summit last year and forthcoming COP29 slated to kick off on November 11 in Baku.
While environmental protection NGOs have criticised the decision to have the oil-rich nations organise the conferences, the accounts have gloated over the initiative.
Nothing new under the sun. Shortly before the launch of COP28 in Dubai last year, dozens of false users cropped up and exuded a refreshing optimism about the role of the Gulf state in promoting climate action.
Now with COP29 around the corner, a wave of hundreds of accounts promoting its Azeri host has washed over the social platform, Northwestern University in Qatar Associate Professor Marc Owen Jones has found.
Yet this time, the tactic is more insidious — boasting more elaborate profiles that claim varied “interests” and credentials ranging from “farmer” and “environmental activist” to “football fan”.
The goal of the campaign “is to legitimise the account, so that it can be picked up by potentially real people, who have an audience and a sounding board”, said Christine Dugoin-Clement, a researcher at IAE Paris-Sorbonne.
‘Recurrent themes’
To seem more “real”, the profiles also tailor the content they publish to the target country — for France, some bots openly slam President Emmanuel Macron’s policies or react to statements by political figures. The posts also target Spain and Germany.
Emirati authorities did not immediately respond to an AFP request for comment.
While its impact is difficult to measure, the suspicious activity grabbed the attention of French authorities, security sources told AFP.
They reported that “the modus operandi in use requires significant financial means that simultaneously allow for the management of numerous accounts, the adaptation of posts to targeted content and countries, as well as the adoption of behaviours to get past the X platform’s moderation policy”, one source said.
Yet clues suggest the posts are not authentic.
Not only did the analysed accounts become active over the summer, despite having been created months earlier, but they also discussed “recurrent themes, some of which appear across multiple languages”, Antibot4Navalny told AFP.
Many used AI-generated images and the same account often reused the same phrases. Words out of place or in another language, such as Chinese characters in a message in French, sometimes slipped through the cracks.
Using AI “reduces the entry cost of this type of operation”, said Dugoin-Clement.
In November 2023, Viginum, a French government agency set up to detect digital disinformation campaigns, linked a campaign to smear the 2024 Paris Olympics to Azerbaijan.
