A MEME FEATURING images generated by artificial intelligence proudly proclaims ‘Ireland Belongs to Them’ below an image of people dressed in 1920s clothing.
Standing next to two adults on an idyllic boreen are two child-sized figures, though their faces are distorted and blurred almost beyond human appearance.
The meme has been shared hundreds of times on a Facebook page where it accumulated more than 131,400 views at the time of writing, according to data provided by Meta, Facebook’s parent’s company.
The image was first posted on 28 November to a page that regularly shares anti-immigration views called ‘Ireland – Rising from the Ashes’.
Another image on the page shows a man in a suit kicking someone so hard they fly through the air, with a caption that reads: “The only way to save Ireland: Mass deportation and close the border”.
However, many of the most popular posts on the page use AI-generated images in attempts to promote racist and anti-immigrant views.
Like the image of the 1920s-esque family, another AI-generated image featuring the slogan “Ireland belongs to the Irish” was posted on the page on the same date. It also features four figures in a rural setting, whose features – including their faces – are blurred and distorted.
Yet another AI-generated image posted to the page on 27 November shows a black woman in a headdress in what appears to be a dusty African village. She is pictured on a motorbike with dozens of children hanging from her and the bike.
“Off to Ireland to reunite with her hubby at the grand canal”, a caption on the image reads. That picture was shared 1,200 times, accumulating more than 260,300 views, according to Meta.
And yet another AI-generated image shows a scared young girl surrounded by older, jeering, dark-skinned men.
Such artificial images are common in anti-immigrant groups and are often used to suggest that white girls will be sexually assaulted by groups of non-white men.
In fact, in the last week, more than 20 obviously AI-generated images were posted to the page to promote anti-immigration or ethno-nationalist views, in addition to memes and crudely edited photos.
These included the AI-generated images of sad-looking old Irish faces, screaming anti-immigration protesters, and a young white man in old-fashioned clothes confronting a black man in a tracksuit.
Much of this content could be categorised as “AI slop”, a term referring to badly made content that can be produced en masse by artificial intelligence and distributed online to see what sticks.
Typically, posts that do well are reposted, or variations on the same theme are generated. In time, these can increase engagement and, in some cases, revenue.
It’s similar to the AI-generated website that duped hundreds of people to gather on Dublin’s O’Connell Street for a non-existent Halloween parade at the end of October.
The use of AI images on ‘Ireland – Rising from the Ashes’ dates back months, long enough ago that the page posted fake images in July celebrating riots in Coolock, which were a response to plans to house asylum seekers in a former Crown Paints factory.
The use of AI-images by anti-immigration groups more generally has increased dramatically in the last year.
The Journal has previously debunked AI-generated images of a homeless Irish family used to push narratives that foreigners “get a place straight away”, stand-offs said to show anti-immigrant protesters facing off against Gardaí, burning buildings in rural Ireland said to have been set alight by “unvetted men”, as well as a deepfake video that appears to show Taoiseach Simon Harris congratulating anti-immigration groups.
“Since the Dublin riots, anti-immigrant groups have been using generative AI tools to create content that espouses their main ideologies – ethno-nationalism, preserving Irish identity and anti-migrant sentiment,” said Ciarán O’Connor, a senior analyst with the anti-extremism think-tank the Institute of Strategic Dialogue.
However, transparency information that features on the Facebook page’s About section shows something interesting: this particular Facebook page was started in 2013 with the name ‘Water metering Ireland – Let’s fight back’.
Initially, the page focussed on opposition to water charges, only veering off occasionally – in one case to promote the spurious claim that fluoride in drinking water is dangerous (a claim now espoused by incoming US President Donald Trump’s pick to lead the Department of Health and Human Services).
However, the creator appeared to be losing interest throughout 2015, with few or no posts for months.
In June 2016, the name of the page changed to ‘Ireland – Rising from the Ashes’ and shortly afterwards, the frequency of posts on the page rose dramatically, though initially the posts focused on politics and other social issues, such as homelessness, protests, and labour strikes.
None of these posts had many interactions.
However, using tools provided by Meta, Facebook’s parent company, The Journal was able to identify more than 40 posts on the page with more than 100,000 views. One is from 2023, the rest are all from 2024.
Of these, seven are clearly AI-generated, one of which has about 640,000 views.
This indicates that the huge volume of posts, many of which are made with the help of AI, may have massively expanded the page’s reach.
The page has made approximately 17,000 posts since being created. More than 6,000 of which were made this year alone, seemingly helped by the use of AI.
“In general, this form of content is easy to produce at scale with little technical proficiency,” Ciarán O’Connor told The Journal, noting that common tropes in far-right groups included stereotypical images of Ireland, fake protests, migrants or non-white people acting suspiciously, or images of Irish people experiencing deprivation.
“It often falls into the uncanny valley, appearing cheap and clearly artificial with obvious spelling mistakes or other inconsistencies,” O’Connor said.
“But, it’s still quite popular among anti-immigrant groups who use this imagery to evoke ethnonationalist sentiments and share them widely across social media platforms.”
The Journal’s FactCheck is a signatory to the International Fact-Checking Network’s Code of Principles. You can read it here. For information on how FactCheck works, what the verdicts mean, and how you can take part, check out our Reader’s Guide here. You can read about the team of editors and reporters who work on the factchecks here.
To embed this post, copy the code below on your site
have your say