Disinformation Campaign on Social Media Reached More Than 40 Million People – but Meta ‘Alarmingly’ Hasn’t Revealed Culprits

A massive disinformation campaign that reached at least 41 million people, pushing anti-immigration and Islamophobic content and fanning sectarian flames in Lebanon for months before the European Elections, has been included in a report this month by Meta.

Facebook’s parent company mentioned the offensive, spread mainly via social media ads, in its second quarterly Adversarial Threat report for 2024, but it was uncovered by Dr Marc Owen Jones and Sohan Dsouza of the open source intelligence community (OSINT). They spent over five months documenting and analysing the campaign which has become known as the Qatar Plot. It was published in July.

The campaign came in two waves, the first from January to March, targeting France, the UK, US and Saudi Arabia, and then, Belgium, France, Sweden, Croatia, Malta, and Germany from May to June, just weeks before the EU elections, the UK General Election on July 5, and not long before the riots erupted in the wake of the Southport stabbings on July 29.

While the Qatar Plot report’s authors had estimated the cost of the campaign to be $100,000 to $250,000, Meta put it at £1.2 million.

In response, the Qatar Plot authors have said they will have to revise their estimates on the reach and impact of the campaign on ballot boxes and street level politics in the UK and US.

Dsouza has said this “has got to be one of the largest – if not one of the top five ever influence operations in Meta’s history”.

The Qatar Plot report couldn’t identify the source of the coordinated campaign although those promoting some of the content include human rights activists (unknowingly), Christian Zionists and far-right accounts.

The campaign attempted to ignite far-right grievances ahead of European Elections from 6-9 June. Some parts, the Qatar report authors noted, were “quite odd”, like physical ads at CPAC, the Republican fundraising event, and in New York’s Times Square.

The Qatar Plot report found the campaign spread its narratives online using at least 978 Facebook ads, hosted on at least 25,000 Facebook Pages, supported by a network of at least 44 “burner” pages, some of which were clearly stolen.

The campaign on X accumulated at least several million impressions/views using several overtly branded accounts supported by more covert advertising/booster accounts, with paid engagement networks likely to have also been used.

Narratives were supported by the vandalisation of at least 47 Wikimedia properties, using nine accounts often working in coordination, and a TikTok video was viewed over a million times, the report notes.

The campaign also created front NGOs and petitions, with Change.Org.

The methods used by the campaign – run via Vietnam – and its success in reaching such large numbers of people while leaving experts unable to identify the culprits raises serious questions around transparency and ability of social media platforms to combat orchestrated disinformation tactics.

The gap between the Qatar Plot report and Meta’s own figures is in part due to the opacity of the Meta Ad Library which also has an issue with disappearing information.

Meta corroborated the Qatar Plot report’s findings and named the same proxy company, “LT Media”, to be the conduit for money funnelled into Meta to fund it. However, it did not offer further information or identify those behind the campaign which Dsouza describes as “alarming, especially given the actual spend”. 

Meta’s report failed to provide the same level of detail as the Qatar Plot report with descriptions of the geo-targeting of advertising only mentioning France, and not the smaller-yet-heavy targeting of other countries.

Meta also focused on the anti-Qatar messaging in much of the operation but left out how a lot of that messaging rode atop anti-immigrant and Muslim content.

It’s also unclear if Meta noted or has taken action against the many algorithm booster account networks that were manufactured off a common template and used to add inauthentic engagement to posts by the campaign’s primary accounts.

The Meta report identified 49 Instagram accounts which the Qatar Plot report didn’t find. Instagram has less searchability and fewer transparency features than Facebook does which hampered researchers.

It would appear that eight years after the Cambridge Analytica scandal which included the illegal scraping of Facebook data and micro-targeted advertising during the UK’s Brexit vote and the 2016 US Presidential Election, Meta and other social media platforms are still vulnerable to coordinated influence operations. This raises concerns of similar campaigns being carried out in the run up to this year’s US Presidential Election.

During the Qatar Plot investigation, the authors observed many vulnerabilities that let the operation’s content run for days to weeks on the platforms, and encountered many obstacles to transparency.

The operation utilised a myriad of tactics across online platforms to reach at least tens of millions across continents with its messages, and even to make it offline into high-visibility ads and personal endorsements.

Online, it was able to take advantage of numerous weaknesses in platform inauthenticity countermeasures and accountability measures, and of numerous gaps in platform transparency to push its narrative with very little hindrance, validating and stoking xenophobia in the process of attacking a state and pushing geopolitical objectives.

It even used generative AI to create propagandistic imagery, and very likely, to narrate its video content. 

The use of untraceable front NGOs and proxy sponsors/accounts also indicate that those behind the operation sought to keep their identities concealed while spending enormous amounts of money to push their messages out.

Meta is yet to respond to a request for comment and has not officially commented on the Qatar Plot report or publicised their Adversarial Threat report in any meaningful way.