Bb5a31f81ClaimAn investigation was requested in Brazil regarding allegations that Meta disproportionately removed pro-Lula posts while amplifying pro-Bolsonaro content during the 2026 electionAgreeDisagree
Bb5a31f81ClaimMeta removed multiple pro-Lula social media posts leading up to the second round of the 2026 Brazilian presidential electionAgreeDisagree
Bb5a31f81ClaimIn the spring of 2019, a 30-person task force at Facebook drafted recommendations regarding political figures inciting ethnic violence against Muslims in IndiaAgreeDisagree
Bb5a31f81ClaimIn 2019, Mark Zuckerberg established a company-wide policy exempting political speech from Facebook moderationAgreeDisagree
Bb5a31f81ClaimIn December 2020, Facebook dissolved its Civic Integrity team following the 2020 United States presidential electionAgreeDisagree
Bb5a31f81ClaimPrior to the 2020 United States presidential election, Facebook required groups generating excessive calls for violence to appoint internal moderatorsAgreeDisagree
Bb5a31f81ClaimFollowing the 2020 United States presidential election, Facebook disabled election-specific safety systems prior to the January 6, 2021 Capitol riotAgreeDisagree
Bb5a31f81ClaimPrior to the 2020 United States presidential election, Facebook disabled commenting features on groups displaying excessive calls for violenceAgreeDisagree
Bb5a31f81ClaimA BBC investigation found that women were being secretly filmed with smart glasses, and the footage was posted on TikTokAgreeDisagree
Bb5a31f81ClaimAn anonymous comment on an online forum for Greek Life accused the fraternity brothers of gang-raping a sophomoreAgreeDisagree
Bb5a31f81ClaimIn the days after the incident, campus rumors described Jane Doe as a 'slut' and a 'whore.'AgreeDisagree
Bb5a31f81ClaimA single platform filtering content risks losing viewership to less strict competitorsAgreeDisagree
Bb5a31f81ClaimDiscord's content moderation prompts for GPT-3.5 were often refused due to the model's built-in guardrailsAgreeDisagree
Bb5a31f81ClaimDiscord's goal was to use LLMs for custom moderation based on individual server norms, which sometimes conflicted with OpenAI's policiesAgreeDisagree
Bb5a31f81ClaimOpenAI gave Discord early access to GPT-3.5 for use cases including content moderationAgreeDisagree
Bb5a31f81ClaimDiscord's content moderation team consisted of over 5,000 outsourced contractors globallyAgreeDisagree
Bb5a31f81ClaimLinkedIn introduced a "Seems like AI slop" button, which was used by over 1 million users in its first two weeks.AgreeDisagree
Bb5a31f81ClaimFollowing the Cambridge Analytica scandal, Facebook established an oversight board modeled on a Supreme Court.AgreeDisagree
Bb5a31f81ClaimFormer Meta executive Arturo Behar testified that the company scrapped safety recommendations from its own research on young users.AgreeDisagree
Bb5a31f81ClaimThe platform uses LLMs to automatically detect harassment in live streams.AgreeDisagree