Protecting Kids Online: Instagram and Reforms for Young Users

QuestionAsks whether 3 hours/day — the top option in Instagram's new parental time-limit tool — is an appropriate amount of daily use for kids.

ResponseMosseri declines to say whether 3 hours/day is healthy, deferring entirely to individual parental judgment regardless of the value chosen.

QuestionAsks Mosseri to confirm an internal document describing the loss of teen users as an "existential threat" to the company.

ResponseMosseri concedes the document/characterization is likely accurate.

QuestionA sustained cross-examination pressing Mosseri to commit to the UK children's code standard in the US, plus concrete enforcement mechanisms — AG lawsuits and private rights of action.

ResponseMosseri says he supports UK-style safety standards in principle, favors an industry body whose standards get policymaker approval and regulator enforcement, floats tying Section 230 protection to compliance, but repeatedly avoids committing to AG enforcement or private rights of action specifically.

QuestionCites the Haugen 32%/6% teen-harm statistics and asks whether Mosseri supports bipartisan legislation to update COPPA to give 13–15 year olds control over their data.

ResponseMosseri disputes the 6% suicide-ideation figure as inaccurate (says it was actually 1%), says he supports federal (not industry) youth-safety regulation in principle, but raises age-verification concerns rather than directly answering "yes or no" on Markey's specific bill.

QuestionAsks whether Mosseri supports a legal right — and specifically a legal mandate — for minors to have their online data expunged.

ResponseMosseri confirms account/data deletion is already available and says he'd support legislation mandating that companies allow it.

QuestionRepeatedly presses for a direct yes/no on whether Mosseri supports legislation banning targeted advertising to children.

ResponseMosseri never gives the yes/no Markey asks for — describes existing ad-targeting restrictions (age/gender/location only, certain ad categories excluded) and says he'd support "something in the direction" of limiting targeting, without endorsing an outright ban.

QuestionCiting Meta's own letter (15,000 human reviewers) and Frances Haugen's view that human review outperforms AI at catching eating-disorder content, asks whether Mosseri will commit to investing more in human review specifically.

ResponseMosseri doesn't commit specifically to more human reviewers, reframes toward investing in both human and technological review generally.

QuestionPresses on how much of Instagram's review staff actually covers each of the 70 languages the platform serves, and whether that leaves moderation gaps.

ResponseMosseri corrects an earlier misstatement (Instagram is used in more languages than the 70 it reviews content in) and doesn't dispute that coverage gaps exist.

QuestionAsks what concrete steps are being taken today on age verification for under-13s, and why parental controls took this long given Meta's prior experience with Messenger Kids.

ResponseMosseri explains the paused Instagram Kids' parental-controls work was repurposed for teens, describes classifier-based age prediction as imperfect, and frames age verification as an unsolved industry-wide problem rather than answering why Meta waited.

QuestionAsks whether Meta's 21 million CSAM-report submissions to NCMEC in 2020 also flag the individual traffickers/exploiters posting that content, not just the images.

ResponseMosseri doesn't know and commits to following up rather than confirming whether trafficker identities are included in NCMEC reports.

QuestionRejects Mosseri's platform-wide "5 in 10,000" prevalence stat as not matching what actually happened on the test 13-year-old account, and asks why Instagram's own recommended-account algorithm led directly to the harmful content shift.

ResponseMosseri accepts personal accountability if a bad recommendation occurred, and announces the Community Standards Enforcement Report will be independently audited by Ernst & Young starting that quarter.

QuestionCites a Tech Transparency Project (TTP) experiment finding minors could reach drug dealers on Instagram in two clicks, and asks why children's accounts can search drug content at all.

ResponseMosseri says drug-selling accounts violate policy and aren't allowed, downplays the TTP finding as anecdotal, and admits he isn't familiar with the specific report cited.

QuestionRe-asks, more insistently, whether Mosseri agrees with the Surgeon General's recommendation to limit social media usage, after an earlier non-answer.

ResponseMosseri again avoids agreeing/disagreeing directly, redirects to parental-controls tools in development.

QuestionDirectly asks whether Instagram's revenue increases with more time spent on the platform.

ResponseMosseri confirms directly — more time on platform means more revenue.

QuestionPresses on whether Meta holds internal data specifically linking platform usage to teen suicide/mental health.

ResponseMosseri confirms Meta does suicide-related research and describes specific interventions (removing suicide-method content, connecting at-risk users to resources, proactive emergency-service outreach) without ever directly confirming or denying internal usage-linked mental health data.

QuestionFrames "Instagram for Kids" itself as a potential "gateway drug" to more platform usage, and asks why it was paused, whether the pause is permanent, and whether Mosseri worries it would hook kids further.

ResponseMosseri defends the original rationale (kids already online younger than the product was designed for), doesn't answer whether the pause is permanent, and again confirms he personally made the pause decision.

QuestionAsks Mosseri to explain what "engagement based ranking" means.

ResponseMosseri describes engagement-based ranking as surfacing the content Instagram predicts a user will find most relevant/interesting, filtered for community-guideline violations.

QuestionAsks whether Instagram employs behavioral scientists and whether Mosseri would be informed if they discovered a negativity bias affecting teenage girls specifically.

ResponseMosseri confirms data scientists exist and says he'd generally expect to be kept informed of important safety findings, without directly committing on the record to being told of a negativity-bias finding specifically.

QuestionPoints out that Instagram's stated age/gender/location-only ad-targeting limit applies to advertisers but not to Instagram's own ML ad-delivery system, and asks whether that's a double standard.

ResponseMosseri confirms Instagram's own delivery system does use in-app teen activity data beyond age/gender/location to target ads, framing it as relevance rather than the same kind of targeting advertisers are restricted from.

QuestionCites Instagram's own Head of Safety and Well-Being naming a specific viewing-duration threshold (~20 minutes) for negative mental-health impact, and asks how that squares with a time-on-platform-driven business model.

ResponseMosseri doesn't dispute the correlation, reframes "more time = more effect" as symmetric (positive or negative), and points to daily-limit tools and in-progress parental controls rather than addressing the business-model tension directly.

QuestionDescribes a specific constituent case — a 14-year-old groomed on Instagram and trafficked across state lines — whose only legal recourse against Instagram is binding arbitration, and asks whether children harmed this severely should have another avenue for redress.

ResponseMosseri calls the story terrifying but never answers whether trafficking victims should be exempt from mandatory arbitration, redirecting to a general call for federal regulation each time Cantwell asks directly.

QuestionCites the WSJ "Facebook Files" 32%/13%/6% statistics and asks whether Mosseri will finally commit to producing the underlying research to the committee.

ResponseMosseri again disputes the widely-cited stats (reframing 32% as "1 in 3 girls who already suffer from body image issues," and the suicide figure as 1% not 6%), and commits only to "doing all I can," citing privacy and data-retention limits rather than a hard yes on producing the research.

QuestionPresses on the specific format (slide deck, PowerPoint) in which the body-image/suicide research was originally presented to Mosseri.

ResponseMosseri confirms the research was presented via PowerPoint-style slides, claims the specific slide Cruz referenced has already been made public, and redirects toward broader industry-wide data access as the more important goal.

QuestionAsks whether Mosseri was personally concerned by the suicidal-ideation finding, whether Instagram ever quantified actual teen-girl deaths linked to the product, and what specific policy changes resulted from the research.

ResponseMosseri says he's concerned but repeatedly deflects the "did you quantify actual deaths" question three separate times, and cites the Restrict feature (anti-harassment) and unspecified "limits" as concrete product changes stemming from the research.