MENÚ
On January 7, Meta founder and CEO, Mark Zuckerberg, announced his company is “going to get rid of fact-checkers and replace them with community notes similar to X, starting in the US.” Zuckerberg claimed that “fact-checkers have just been too politically biased and have destroyed more trust than they've created.”
Lee esta historia en español haciendo clic aquí.
Zuckerberg’s statement was immediately fact-checked, most notably by the International Fact-Checking Network (IFCN), European Standard Fact-Checking Network (ESFCN), and LatamChequea network among others. In an open letter signed by more than 125 fact-checking organizations from around the world, the IFCN called Zuckerberg’s comments “false.” (Two of us represent organizations that signed the letter.) Questioning his motivations and the specifics of his claims, the IFCN concluded that “the decision to end Meta’s third-party fact-checking program is a step backward for those who want to see an internet that prioritizes accurate and trustworthy information.”
That observation is born out by the results of a report we released today titled “Platform Response to Disinformation during the US Election 2024.” The report contains the results of an analysis of the responses of very large online platforms to debunked disinformation in Spanish during the 2024 US Election.
What we found
As in prior elections, in 2024, Spanish-speaking communities were again the specific targets of electoral disinformation efforts. Importantly, Hispanic Americans are more likely than Black or White Americans to favor social media as a source of news, according to the Pew Research Center, and social media serves as an important channel for disinformation campaigns targeting them.
These communities have not only been targets of disinformation campaigns but have also been at the center of disinformation narratives spreading in the US. In the 2024 electoral cycle, hoaxes falsely linked Hispanic communities to crimes, pushed disinformation on migration policies and statuses, and claimed that undocumented migrants were arriving to register and vote, committing fraud. Recurrent narratives and hoaxes also targeted the main candidates in this election. 49% of the disinformation contained in the posts evaluated in this investigation targeted the main presidential candidates, Republican Donald Trump and Democrat Kamala Harris, and their vice presidential running mates.
In order to better understand how big tech firms addressed such disinformation, we evaluated how five of the biggest online platforms (Facebook, Instagram, TikTok, X, and YouTube) responded to disinformation content aimed at Latino audiences. For that purpose, we analyzed the disinformation content debunked by IFCN-certified US-based fact-checkers in Spanish during the four months ahead of the elections, including Election Day (from July 14 to November 5, 2024).
More than half of the disinformation debunked did not receive any visible action. In total, 45% of posts containing disinformation received visible actions from the platforms; however, the reaction rates and specific content moderation decisions varied throughout the platforms. Facebook is the platform showing the highest proportion of visible actions (74%).
Instagram ranks second, accounting for 59% of observable action, of which 95% constituted debunk labels, the highest percentage among all platforms.
TikTok's proportion of visible actions (32%) surpassed that of X and YouTube. TikTok is the platform that most frequently removed content (79% of its actions). X and YouTube are the platforms with the lower percentage of disinformation content with visible actions, 24% on X and 19% on YouTube.
Community Notes represent 46% of actions on X while contributing to only 12% of the identified disinformation content. Among the 20 most viral posts or videos that received no action, 19 were hosted on X, each of them accumulating over 6.5 million views.
Overall, platforms responded similarly to disinformation content posted in Spanish and in English. However, Facebook posts in Spanish had visible actions 79.7% of the time vs. 45.8% for posts in English), while X took more visible actions on English-written content (28.3%) than in Spanish (15.2%).
Disinformation targeting candidates was the most prominent from the posts strictly related to the most analyzed topic, ‘Elections,’ while ‘Migration’ was the second topic, accounting for 19% of the disinformation in the sample.
Meta’s retreat from fact-checking will likely have significant implications in election periods
Contrary to what was expected, across all of the platforms we studied, disinformation in Spanish received on average 19.7% more visible actions across platforms than in English. The database gathered posts written in both Spanish and English with claims relevant to the Latino community that were previously debunked in Spanish by US-based fact-checkers. However, this figure is heavily influenced by the high number of actions on Spanish-language content on Facebook (33.85% more than those in English on the platform).
These results seem to indicate the relative success of Meta’s efforts to take action on false claims in Spanish that were flagged by fact-checkers in comparison with the other services. The company’s decision to move to a “community notes”-style approach will likely result in far more disinformation going unaddressed, at least according to our observations in 2024.
It is also notable that all platforms performed slightly worse than in a similar study conducted on data collected during last year’s European Union parliamentary elections. In comparison, the US election disinformation analyzed received 12% less visible actions than disinformation around the June 2024 EU elections. It is possible that the mere existence of the Digital Services Act and the active scrutiny of regulators produced this disparate result.
Ultimately, Meta’s decision to abandon independent fact-checking could not come at a more precarious moment. Our analysis shows the enduring prevalence and perniciousness of disinformation targeting Latino communities. While hardly perfect, Meta’s platforms previously demonstrated a capacity to address such content–a capacity it should build, not dissolve. Without robust, proactive intervention by fact-checkers combined with other content moderation tools and systems, the integrity of information–in particular for vulnerable communities–remains at risk.
Read the report in English or in Spanish.
Factchequeado is a verification media outlet built by a Spanish-speaking community to tackle disinformation in the United States Do you want to be part of it? Join us and verify the content you receive by sending it to our WhatsApp +16468736087 or to factchequeado.com/whatsapp.
