Disinformation Nation: Social Media’s Role in Promoting Extremism and Misinformation (QFR Responses)

QuestionNotes research shows emotional and misinformation content is more engaging, and asks how that squares with claims users don't want it.

ResponseSays pushing extreme content isn't in Facebook's long-term interest since unhappy or angry users are less likely to return.

QuestionSchakowsky asks how Facebook will create recourse channels to prevent systematic over-flagging of benign minority-community content (citing the Sikh community specifically) as hate speech.

ResponseFacebook acknowledges it "sometimes incorrectly" flags benign minority-community content as hate speech, and points to its general appeals process ("Request Review"/"Disagree with Decision") rather than any Sikh-community-specific fix.

QuestionCiting that over 15% of youth social media users experience cyberbullying, with minority youth receiving disproportionately more vitriolic messages, Schakowsky asks what recourse Facebook provides at the intersection of cyberbullying and racial/ethnic hate speech.

ResponseFacebook confirms its Bullying and Harassment and Hate Speech policies apply, and specifically states it provides "heightened protections for users between the ages of 13 and 18" because bullying can have more emotional impact on minors, without addressing the specific racial-disparity data cited in the question.

QuestionSchakowsky asks Zuckerberg to detail whether Facebook's algorithms and recommendation tools amplified dangerous/conspiracy content, specifically asking for a case-by-case account of Capitol riot suspects' group memberships, ad targeting, and whether Facebook's recommendations connected them to QAnon or other rioters.

ResponseFacebook declines entirely to answer the specific Capitol-rioter case questions (group membership, ad targeting, recommendation history), citing the ongoing law enforcement investigation, and instead describes its general Dangerous Organizations and Individuals policy and QAnon/militia enforcement actions.

QuestionCiting the Kenosha shooting, Schakowsky asks what Facebook will do to prevent pages calling for armed militia organizing, and how it will enforce that.

ResponseFacebook cites an August 2020 framework flagging locations at elevated risk of armed violence/intimidation (treating them like schools or polling places) and its broader militarized-social-movements ban, without directly addressing page-creation prevention.

QuestionSchakowsky asserts Facebook is failing to enforce its ban on microtargeting ads to people with disabilities, resulting in children being harmed and scammers monetizing fake cures, and asks what Facebook will commit to.

ResponseFacebook says it does not offer disability/medical-condition ad targeting and has been removing such interest categories over time, without directly addressing the specific allegation that children were harmed by microtargeted fake-cure ads.

QuestionSchakowsky asks how much ad revenue Facebook generated from QAnon before banning it, whether Facebook does financial modeling of user-loss risk when making content-moderation decisions about militarized social movements, and whether it will share that internal analysis.

ResponseFacebook does not provide a QAnon-specific ad-revenue figure or share any financial modeling/user-loss analysis, instead asserting broadly that there is "no incentive" to keep hateful content since users and advertisers don't want to see it.

QuestionSchakowsky presses Zuckerberg to clarify whether his repeated "99%" removal-rate claims for terrorist propaganda, human trafficking, and CSAM content mean 99% of what Facebook catches, versus 99% of all such content that actually exists — noting a study finding the real removal rate may be only 25-30%.

ResponseFacebook does not directly reconcile its "99%" removal claim against the cited 25-30% study, instead pivoting to its "prevalence" methodology and reporting specific Q4 2020 upper-limit prevalence figures: terrorism content at no more than 0.07% of views, and "child nudity and sexual exploitation of children" content at no more than 0.05% of views.

QuestionCiting former Facebook policy employees who said Zuckerberg personally intervened to change the rules on banning Alex Jones and the Oath Keepers because he "didn't like the punishment," Schakowsky asks what role Zuckerberg personally plays in content-removal decisions.

ResponseFacebook confirms both bans eventually happened and that Zuckerberg is "ultimately responsible" for major decisions as company leader, but does not directly address or deny the specific allegation that he personally intervened to soften an earlier punishment for Alex Jones.

QuestionCiting 2016 Facebook research finding 64% of people who joined extremist Facebook groups did so via Facebook's own recommendation tools, and Zuckerberg's earlier Senate-hearing non-answer ("I'm not familiar with that specific study"), Schakowsky asks whether he's now familiar with it and whether extremist-group recruitment via recommendations declined after policy changes.

ResponseFacebook characterizes the 64% statistic as tied to a narrower 2017 Germany-specific issue that was "addressed at the time," denies ever having "buried research," but does not directly state whether extremist-group recruitment via Facebook's recommendation tools declined since the policy changes, as specifically asked.

QuestionSchakowsky asks why and how the #Sikh and #Sikhism hashtags were repeatedly blocked in 2020, and whether the blocks were automated or human-driven.

ResponseFacebook confirms the block was caused by a human review error, took nearly three months to catch (March 7 to June 3, 2020), and apologizes, without detailing what specific process changes were made to prevent recurrence.

QuestionSchakowsky asks whether white supremacist/racist hate content is inherently harder to moderate than COVID/vaccine misinformation, given Facebook touted success on the latter while the former "runs rampant."

ResponseFacebook does not directly answer whether hate content is "inherently more difficult" to police than health misinformation, instead citing general hate-speech enforcement statistics (97%+ pre-report detection, 0.07-0.08% prevalence in Q4 2020).

QuestionCiting the 2018 Fair Housing Act allegations that Facebook's ad-targeting systems let advertisers exclude audiences by race, gender, and zip code, Clarke asks whether Facebook still maintains its Section 230 civil-rights-immunity legal position, and what it's doing about discriminatory ad targeting generally.

ResponseFacebook does not directly answer whether it still holds its Section 230 civil-rights-immunity legal position, but describes having eliminated age/gender/zip-code targeting for housing, employment, and credit ads specifically, following settlements with civil rights groups.

QuestionCiting an ADL survey showing only 14% of reported physical-threat content actually gets deleted (down from 22% the year before), Kelly asks why Facebook isn't investing more in moderation and protecting harassment targets.

ResponseFacebook cites its rising bullying/harassment enforcement volume (3.5M to 6.3M pieces of content quarter-over-quarter) and proactive-detection-rate improvement (26.4% to 48.8%), without directly addressing the specific 14%-of-physical-threats-removed statistic from the ADL survey.

QuestionRice asks how Facebook's content-moderation methods differ between misinformation, disinformation, and hate speech (following its policy shift reclassifying Holocaust denial as hate speech rather than misinformation), and how it should be held accountable given an ADL study still finding significant Holocaust denial content on the platform.

ResponseFacebook explains its terminology distinctions (misinformation vs. disinformation vs. hate speech) and cites general hate-speech enforcement statistics, but does not directly address the ADL's finding that significant Holocaust denial content persists despite the policy change, nor propose a specific accountability mechanism.

QuestionCiting the pattern of YouTube Kids users migrating to the main YouTube platform, Trahan asks whether Messenger Kids usage has led to an increase in under-13 monthly active users across Facebook's other apps.

ResponseFacebook denies Messenger Kids automatically feeds into a Facebook account at 13, describes its age-verification approach (self-reported birthdate, user reports, AI age-estimation tools), and argues that kids-specific apps like Messenger Kids actually reduce the incentive for children to lie about their age to access the main platform — without directly answering whether under-13 monthly active users increased.

QuestionCiting leaked 2017 internal documents describing Facebook's ability to monitor young people's posts and photos in real time to detect emotional states like feeling "stressed," "worthless," "useless," and a "failure," Trahan asks how many projects have gone through Facebook's subsequently-announced "enhanced review" process for sensitive/age-specific research, and whether that process has been independently audited.

ResponseFacebook says the 2017 "insecure/worthless teens" research violated its own internal review processes at the time and was never used for ad targeting, describes its post-2017 enhanced-review framework in general terms, but does not answer either specific question asked — how many projects have gone through enhanced review since 2017, or whether that process has been independently audited.