The Digital Battleground: How Online Platforms Failed to Police Election Lies
In the high-stakes arena of a presidential election, the integrity of the information voters consume is paramount. Yet, a sobering new report from the European Digital Media Observatory (EDMO) reveals a troubling reality: the digital gatekeepers entrusted with curbing online falsehoods largely stood by as misleading content flooded social media during Ireland’s 2024 presidential race. The report paints a picture of a system under strain, where the automated and human review processes of major tech companies consistently failed to remove or label the very content their own policies claim to prohibit. This inaction raises profound questions about the effectiveness of self-regulation and the platforms’ commitment to protecting democratic processes, suggesting that the battleground for public opinion is being ceded to malicious actors who exploit the chasm between corporate policy and practical enforcement.
EDMO’s investigation, carried out by its Irish hub during the critical period between late September and early November, involved monitoring a digital ecosystem teeming with problematic material. Researchers identified and formally flagged 117 distinct pieces of content to the major social media giants—a process designed as a “rapid response” mechanism to ensure time-sensitive electoral misinformation is reviewed promptly. The findings, however, are stark. Of the 85 items specifically categorized as outright disinformation—which included maliciously fabricated quotes from public officials, systematic allegations of election fraud, and AI-generated impersonations of candidates—platforms took any form of enforcement action on a mere 13 of them. This near-total inaction reveals a fundamental disconnect between the stated commitment of companies like Meta, YouTube, and TikTok to combat election interference and their operational reality on the ground.
This pattern of neglect extended beyond simple falsehoods to more tangible harms. The report detailed how the platforms’ response to harassment and abuse, particularly videos depicting the intimidation of political figures and party canvassers, was also woefully inadequate. Of the flagged incidents involving threats and abuse on the streets, action was taken in only nine cases out of a larger pool of flagged content. While the report notes that platforms were slightly more responsive to illegal content like videos of spoiled ballots, the consistent failure to address the other categories creates a permissive environment where disinformation can thrive unchallenged. The research also highlighted a further dimension of the problem: much of the flagged material originated from monetized accounts or those that had previously been banned, suggesting that bad actors are not only undeterred by platform policies but are often finding ways to profit from their spread of falsehoods.
The disparity in how different platforms responded to the flagged content is perhaps the most damning finding of all. The report’s data shows a clear hierarchy of inaction, with Meta’s properties, despite receiving the most flags, demonstrating a lackluster response. Instagram, which is owned by Meta, acted on the highest percentage of its reported items at 59%, while its sister platform, Facebook, which had the highest number of flagged items at 54, took action on just 44% of them. TikTok acted on 54% of items, and YouTube on 48%. The professional networking site LinkedIn, also owned by Microsoft, took no action on any of the items that were flagged to it. This inconsistency suggests that while some platforms have honed more effective review processes, others are either unwilling or structurally incapable of enforcing their own standards, leaving users’ information security dependent on the arbitrary whims of which social network they choose to use.
A central thread in the disinformation narrative was the campaign of former mixed martial arts fighter Conor McGregor. The report notes that McGregor “advanced a narrative that his failure to secure a presidential nomination would constitute a betrayal of democracy and the will of the Irish people.” This powerful, emotionally charged claim, which framed his defeat as an establishment conspiracy, did not fade from the public discourse when his campaign ended. Instead, the report found that elements of this narrative were adopted by other unsuccessful candidates, most notably in the form of campaigns encouraging the public to spoil their ballots as a form of protest. While the report is careful to acknowledge that spoiling a ballot is a legitimate form of democratic protest, it warns that this protest—when coupled with false claims about the legitimacy of the nomination process—can undermine trust in the entire electoral system itself, transforming a valid expression of dissent into a tool for subverting democratic confidence.
Ultimately, EDMO’s report serves as a damning indictment of the current state of platform governance. The conclusion that platforms only acted on a fraction of the disinformation that was proactively flagged to them by expert researchers suggests that their much-publicized efforts to police elections are often performative rather than substantive. The fact that action, when taken, was generally swift—often within 24 hours—proves that the infrastructure for enforcement exists. The failure, then, lies not in technological capability, but in a lack of political will or a deliberate cost-benefit analysis where the financial and reputational risk of leaving content up is considered lower than the effort required to moderate it. As the EDMO’s “Democracy Shield” initiative suggests, the fight against online disinformation is an ongoing one, but this report demonstrates that the tech industry’s side of the shield is riddled with holes, leaving democratic processes dangerously exposed to the corrosive influence of unchecked digital falsehoods.