Social Media Accounts Posing as Irish Investigated Amid Protests, Dáil Motion
An investigation this week explored potentially inauthentic social media accounts posing as Irish amid political tensions. Accounts with suspicious posting styles and locations were scrutinized, revealing some discrepancies. The investigation highlights the difficulty in verifying online identities and the unreliability of social media, even with transparency tools.
Peter Steiner's 1993 New Yorker cartoon, "On the internet, nobody knows you’re a dog," remains relevant in the social media age. This week, while scrolling through social media posts from profiles presenting themselves as Irish, the author investigated potentially inauthentic accounts amid fuel price protests and a Dáil motion of no confidence.
The investigation involved identifying accounts with suspicious posting styles, such as those claiming "here in Ireland," repeating talking points, or using Americanized spellings. Accounts with tricolours, AI-generated images, or stereotypically Irish handles were also scrutinized. X's transparency information was checked to verify account locations. Some accounts, like "National Review" and "Cú Chulainn," were found to be posting from West Asia and Brazil, respectively, despite claiming to be in Ireland. A US-based Facebook account, "Real Irish News," shared false information about Taoiseach Micheál Martin.
Despite suspicions based on posting styles, X indicated that many of these accounts were posting from Ireland. This created uncertainty, as Irish users could genuinely write in American English or use unconventional punctuation. Even VPN usage didn't provide conclusive proof. The author's suspicions were based on confirmation bias, a dangerous position for journalists.
The investigation highlighted the unreliability of social media and the difficulty in trusting online identities. The advent of LLMs has increased questioning of content authorship. Tools like location tags and account transparency features are easily manipulated, leading to ambient uncertainty. This makes it harder to discern representative, coordinated, or simply noisy content, fostering the belief that anyone online could be a "dog."