Protecting Kids Online: Facebook Instagram and Mental Health Harms

QuestionCiting an internal document calling tweens "a valuable but untapped audience," Klobuchar asks why 10-12 year olds are valuable to Facebook, what the lifetime value of an under-13 user is, and for copies of any internal financial-value analyses of young users presented to officers/directors.

ResponseFacebook doesn't answer why tweens specifically are valuable beyond generic "understanding how teens use technology"; says it doesn't calculate lifetime value for (the nominally-forbidden) under-13 Messenger Kids users; and declines to provide the requested internal financial-value analyses or say whether any were presented to officers/directors.

QuestionKlobuchar asks for the Facebook-specific (not just Instagram) under-13 account removal count, and who is ultimately responsible for under-13 account-removal policy.

ResponseFacebook gives the 2.6M Facebook / 850K Instagram removal figures and describes its age-estimation AI methodology (including the "Happy 21st Bday" signal-detection approach); does not name a single individual with ultimate responsibility, instead describing a distributed policy team/process.

QuestionCiting a specific reported example (a 13-year-old test account being algorithmically fed pro-eating-disorder accounts like "I have to be thin"), Klobuchar asks how the recommendation algorithm produces that result, what's been done to remove such accounts, who's ultimately responsible, and what specific actions followed the news reports.

ResponseFacebook doesn't directly explain the algorithmic mechanism that surfaced the pro-eating-disorder accounts to the test profile; describes its removal statistics, reporting tools, and the "nudges"/"Take a Break" features; names no specific individual as ultimately responsible, again pointing to a distributed team.

QuestionKlobuchar presses for specifics on what Davis personally did in response to the internal research, who helped develop any recommendations, whether they were enacted, who was responsible, whether any recommended actions were declined, and whether anyone verified the changes actually worked.

ResponseFacebook answers Q2 with a general list of product features attributed to "our research" broadly, then answers every subsequent specific question (Did she personally recommend changes? Who helped her? Were any declined? Was follow-up verification done?) by referring back to that same general answer — never naming Davis's individual role, any specific person who helped her, or confirming any verification/follow-up occurred.

QuestionCiting whistleblower Frances Haugen's testimony that underage users are on the platform "in far greater numbers than anyone is aware," Luján asks whether Facebook will publish its underage-detection methodology.

ResponseFacebook explicitly refuses to publish the detection methodology, stating outright that detailed disclosure would give evaders a roadmap — a direct, reasoned "no" rather than a dodge.

QuestionLee confronts Davis with a specific third-party experiment (Technology Transparency Project) showing Facebook approved ads for alcohol, prescription drugs, vaping, eating-disorder tips, and dating profiles targeted at 9.1 million 13-17-year-olds — directly contradicting her live testimony that such ads aren't allowed.

ResponseFacebook concedes the ads violated its own policies and says they "ultimately did not run" (because TTP itself stopped them, not because Facebook caught them), announces a policy change restricting under-18 ad targeting to age/gender/location only going forward.

QuestionCiting Frances Haugen's testimony that Facebook's AI "only catch[es] a very tiny minority of offending content," Lee asks Davis to confirm or deny this and explain the ad-approval AI process.

ResponseDavis doesn't directly confirm or deny Haugen's "tiny minority" characterization; acknowledges enforcement "isn't perfect" and describes the automated-plus-human review process in general terms.

QuestionLee asks whether Facebook collects interest-category data on teens for adult products (tobacco, alcohol, drugs, sexual content) even when not currently serving ads against that data.

ResponseDavis doesn't give a direct yes/no on whether interest-category data (including adult-product interests) is still collected on teens even without being used for ad targeting — describes general data categories and the new under-18 ad-targeting restriction, which addresses targeting but not underlying collection.

QuestionCiting a WSJ report that an outside consultant recommended Facebook adopt a policy to not profit from trafficking-related ad revenue, Lee asks whether that policy was adopted and, if not, why not.

ResponseDavis doesn't directly confirm whether the specific "don't profit from trafficking ad revenue" policy the consultant recommended was adopted — describes general anti-trafficking policy consolidation, enforcement staffing, and support for SESTA/FOSTA legislation instead.

QuestionBlackburn follows up on her live-testimony question, asking Facebook to describe its parental-consent process for research and to produce a copy of the consent form.

ResponseFacebook describes its consent process and, notably, does actually attach the requested consent form (as an exhibit to the QFR response) — unlike most other document/form requests in this hearing, which were deflected.

QuestionBlackburn asks how Facebook recruits child/teen research participants and whether it employs child psychologists or therapists for this work.

ResponseFacebook confirms it does employ researchers with clinical/developmental psychology and pediatric backgrounds (among other fields), and describes email/in-app recruitment plus a third-party parent panel for Messenger Kids research.

QuestionBlackburn asks whether data from removed underage accounts is actually deleted or retained in case of future reactivation.

ResponseFacebook says data is deleted per standard policy once an account is permanently disabled (30 days after failing an age checkpoint), though doesn't elaborate on what "standard deletion policies" actually entail (retention period, backups, etc.).

QuestionBlackburn names the specific internal report ("Apple Escalation on Domestic Servitude—how we made it through this SEV") and presses Davis to explain, in detail, why the report's own words ("absence of proactive detection") describe exactly what Davis disputed live.

ResponseFacebook does not explain "in detail what happened in this case" as asked, and doesn't address the "absence of proactive detection" language from its own report — pivots to a general description of anti-trafficking program investments instead.

QuestionCiting Frances Haugen's testimony and data shared with her office, Blackburn presses on Facebook's use of "MAPMAPMonthly Active PeopleCore Meta engagement metric counting distinct people active on a platform in a month. Paired with DAP in internal goal-setting; the MAP/DAP ratio is used as a measure of how frequently a monthly user returns.Full entry ↗" (monthly active people) to estimate the underage population — directly following up on the live-testimony "map" sub-thread from Blackburn's first round.

ResponseFacebook doesn't confirm or deny using MAPMAPMonthly Active PeopleCore Meta engagement metric counting distinct people active on a platform in a month. Paired with DAP in internal goal-setting; the MAP/DAP ratio is used as a measure of how frequently a monthly user returns.Full entry ↗ specifically to estimate the underage population (the actual question asked), instead describing its general age-detection/account-removal methodology.