How Iran outdid SNP in promoting Scottish independence on social media
Iran’s Islamic Revolutionary Guard Corps has been waging a social media campaign in support of independence, while AI bots have been attacking John Swinney
In December of last year, an X account claiming to be written by a Scottish man posted the slogan: “Yes to Scotland’s future. No to a future decided elsewhere.” The post included the hashtag #ScottishIndependence and an image of a crowd waving Scottish flags.
The message of the post was clear to most users who saw or shared it. The irony, however, was not. The person behind the X account wasn’t from Scotland at all, but rather somewhere decidedly “elsewhere”.
Like many politically engaged accounts on X, he posted routinely, often criticising the British government and loudly advocating for the Scottish National Party and independence.
He was highly likely part of a social media influence campaign waged by or for Iran’s Islamic Revolutionary Guard Corps that has been targeting conversations around Scottish independence since at least 2021. Their goal is to further divide the United Kingdom.
READ MORE: A democratic transition government for Iran already exists. It's time to end 47 years of tyranny
Foreign trolls and bots
Our team at Clemson University’s Media Forensics Hub tracks online foreign information manipulation and interference to better understand their tactics, goals, and impact. We’ve built tools to identify inauthentic or coordinated attempts to influence social media conversations, separating foreign troll and bot needles from the real human haystack.
We first identified this Iranian campaign in 2024. At its peak, Iran operated at least 100 accounts purporting to be from the UK, and these accounts combined had a quarter of million followers. Our analysis showed that, for a time, the Iranians generated more engagement on X around the issue of Scottish independence than the SNP itself.
Since then, consistent tracking and reporting on the Iranian campaign from various teams has effectively pressured X to periodically take action and suspend accounts. The X account highlighted above was part of what was at least the third generation of Iranian assets created to be part of this operation.
Abrupt switch to Iranian propaganda
We linked him to the campaign through a combination of behavioural signals and account meta data. He was one of several dozen Iranian accounts we reported on last week that were active across X, BlueSky, and Instagram, all of which have been suspended by the platforms.
The entire network, which also included Irish accounts calling for unification, had transitioned from talking about national politics to focusing exclusively on Iranian war propaganda. Using new accounts and lacking organic followers, the campaign has lost the influence it once had.
It nonetheless points to an important strategy of modern state-backed influence operations: identify your adversary’s pain point and press until it hurts. Scottish independence is clearly seen by Iran and other state actors as a vehicle to harm the UK and the Western order through means short of kinetic warfare.
SNP under attack from AI bots
The digital battlefield is not one sided. The comment threads on X of posts from First Minister John Swinney and members of the SNP show hundreds of critical comments from another, distinct, set of profiles. These comments originate from accounts with profile images that typically show attractive women whose online interests seem to be exclusively centred around crypto currencies and Scottish politics.
Their replies are AI generated, occasionally incorporating a comically tone-deaf Scottish brogue, demonstrating the need for subtler prompts from their operators. In one angry response to Swinney, for instance, a user posted: “Aye, ye’ll nae believe the gall o this wee SNP toad, John bloody Swinney.”
These accounts are operated by a prolific dark marketing company paid to inauthentically amplify clients’ content across platforms. Someone is paying to harass the current Scottish Government, illustrating the legitimacy of concern regarding the impact of AI on political discourse.
Most accounts are real but...
We should, of course, be careful not to overly inflate the present risks. The great majority of accounts, including those discussing divisive political issues, are real people. In a recent piece in The Times, former Foreign Secretary William Hague cited the disinformation research firm Cyabra in suggesting that 26 per cent of accounts arguing for Scottish independence on X were fake.
According to our own research, this number is unrealistically high, wrong by an order of magnitude. Most accounts are genuine (especially the ones that disagree with you!), but differentiating what is real from what is not is difficult.
There is no question, however, that motivated actors are trying to control online narratives, including those around Scottish politics, and AI is making their job cheaper and easier. Experts are warning the problems posed by fake social media accounts may soon be worse than Cyabra suggests.
A recent article published in the journal Science warned: “In today’s fragmented information environment, ideological echo chambers offer fertile ground for manipulation. AI swarms are distinctly equipped to exploit this by engineering a synthetic consensus that appears to bridge these divides.” Bandwagon appeals, in other words, are an effective and persuasive strategy, and AI bots are the perfect tool for making them.
Defending democracy
Defending against inauthentic manipulation will be a continuing contest between evasion and detection. The first step in addressing any problem is ensuring the issue has been accurately sized and understood.
As our research has consistently demonstrated, the technology and methodologies required to do this at speed already exist. It is therefore essential that better tools and approaches are deployed to safeguard democracies from malign influence.
The involvement of foreign and inauthentic actors in domestic political discourse is unlikely to diminish and has in fact accelerated in recent years. Addressing this challenge will require a proactive and carefully considered strategy, particularly as the approaches used to date appear to have had little impact in deterring continued manipulation.
Professor Darren Linvill and Professor Patrick Warren are co-directors of the Clemson University Media Forensics Hub in South Carolina, USA
Comments
Want to join the conversation? Please or to comment on this article.