Online Platforms Failed to Remove Most Disinformation During 2025 Election, Report Finds

Most social media posts identified as electoral disinformation during the 2025 presidential election were left online, according to a new report from the European Digital Media Observatory (EDMO). The EDMO, which works to promote research on disinformation across 27 countries, found instances of content spreading false claims about the nomination process on platforms like Meta, YouTube, and TikTok. While these platforms participate in a "rapid response" system for flagged content, they are not obligated to act.

EDMO Ireland monitored online content between September 24 and November 4 last year, flagging 117 items. The report noted that many flagged posts were linked to monetized accounts or those previously banned, suggesting a pattern of problematic activity. The researchers estimate the total number of such posts was likely much higher than what they flagged.

Of the 117 items, 85 were marked as disinformation, 26 were AI-generated, 49 were illegal election-day content, and two related to political advertising. Specific examples included fake quotes from officials, claims of systemic fraud, AI impersonations of candidates, and videos of harassment.

Platforms took action on only 13 of the 85 disinformation posts, primarily Instagram and TikTok. Action was also taken on nine of the 13 posts flagged for harassment. Facebook had the highest number of flagged items (54), but Instagram was the most responsive, acting on 59% of reported content, followed by TikTok (54%), YouTube (48%), and Facebook (44%). LinkedIn took no action on any reported items.