highbeamSubscribe

Policy brief 03 · Kids + Social Media

Content Ratings + User Controls

Give families the dials advertisers already have.

All policy briefs

On YouTube, an advertiser can keep its ads away from strong profanity, dramatized violence, and sexually suggestive themes.1 A viewer gets Restricted Mode: one on-off switch for "potentially mature content," set browser by browser.2 Congress should require large platforms to offer families the controls they give advertisers, rate recommended content on five scales that users and parents set, and prove the filters work.

The Problem

Roughly nine in ten American teens use YouTube, and about six in ten use Instagram and TikTok.3 Families disagree about what a 12-year-old should see, and about what they want in their own feeds. A single company policy cannot settle those questions for all of them.

The platforms already rate content. Instagram now places teens in a "13+" setting "inspired by movie ratings criteria and parent feedback," with a stricter option parents can choose.4 But the scale is the company's: Instagram offers levels, not a separate dial for violence, language, or sexual content.

Other mass media hand families that dial. Films have carried ratings since 1968 and video games since 1994;56 since 2000, TV sets 13 inches or larger have had to include a V-chip that blocks programs by rating.7 Social feeds have no common rating a family can set. Three gaps leave families without a handle:

  1. Advertisers get the dials. Meta lets advertisers choose among three sensitivity filters, exclude topics such as news and politics, block specific content, and review delivery reports on the individual pieces of content their ads appeared beside.8
  2. The ratings serve the company. YouTube has used classifiers since 2011 to identify "racy or violent" videos and keep them out of recommendations,9 on the company's own terms.
  3. Parents face two bad options. Say no, and a child loses the place where 74% of teens say they feel more connected to friends;10 say yes, and the platform's settings, rather than the family's values, decide what fills the feed.

Why legislation: Congress has done this before. In 1996 it found that giving parents tools to block "violent, sexual, or other programming" was "nonintrusive and narrowly tailored," required the V-chip, and left the ratings to industry, specifying that nothing in the law authorized rating programs for their "political or religious content."7 Courts have since pointed to user-chosen content filters as a less restrictive path than speech mandates.11 No federal law gives users or parents category controls over a recommendation feed. When a company decides what millions of children see, the family should hold the dial.

The Solution

A four-step staircase: each step stands alone, and each step up gives families finer control and stronger guarantees. Scope: social and video platforms whose recommender systems average more than 10 million monthly U.S. users; the controls govern what a recommender chooses to show, not what a user searches for.

Step 1 — Give families the advertiser's dials. Any platform that lets advertisers avoid categories of content must let users, and parents of minors, filter the same categories out of their own recommendations, using the same classifications. It asks platforms only to open tools they already built for advertisers.

Step 2 — Rate content; let users set limits. Before recommending text, images, or video, platforms rate it on five scales: violence and extreme harm; profanity and language; sexuality and sensuality; self-harm or body image; and sensationalism and other defined themes. Each scale has intensity levels and room for context such as news, education, or recovery. Users choose a threshold on each scale, including none, and anything above it is excluded before delivery rather than merely ranked lower.

Step 3 — Put parents in charge. Verified parents set and lock the thresholds for users under 18, with notice to the child, a process for disputed custody, and no access to private messages. Under the House-passed KIDS Act, a parent could view a 15-year-old's settings but change them only for children under 13;12 here, the parent decides.

Step 4 — Prove the filters work. Independent audits test missed content and overblocking across formats and languages; platforms publish error rates and let creators contest a rating. Families may also delegate the job to third-party filters through secure interfaces, as Sammy's Law proposes.13

Where to start: Step 1 is the floor; it asks platforms to share controls advertisers already use. Step 2 is the heart, and Step 3 puts it in parents' hands.

Administration and enforcement: The FTC sets functional standards within 12 months, with compliance a year later. Ratings describe content and context, never a speaker's identity, politics, or religion. Civil penalties for failing to offer controls or to honor chosen settings; liability turns on reasonable implementation and honest claims, since no classifier is perfect.

Risks and Mitigations

  • Government control of speech: Users choose the limits, and ratings ignore viewpoint, as Congress specified for television in 1996. A court could still treat a mandated rating as compelled speech, and "sensationalism" is the most subjective scale; it needs narrow, concrete criteria. The risk remains.
  • Overblocking: Classifiers err, and Meta itself cautions that its content moderation systems are "not the same as a movie ratings board."4 Protect health, recovery, and news contexts, publish error rates, and offer appeals; some mistakes will persist.
  • Parental tools misused: Settings govern recommendations only. Parents see thresholds, never messages, and the rules include a process for abuse and disputed custody.

Similar Bills

Fit measures similarity to this proposal's mechanisms: High = direct precedent; Partial = useful component with material differences; Related = adjacent approach.

Federal — 119th Congress

Proposal or bill Relevant provisions and fit Fit
H.R. 7757 — KIDS Act
Guthrie (R-KY)
Passed House · June 29, 2026		 §214(a)(1)(F) gives minors prominent options to opt out of personalized recommendations and to limit "types or categories" of recommendations; §214(a)(3) makes the most protective setting the default. Closest precedent for Step 2; no rating scales or intensity thresholds, and parents may change settings only for children under 13. High
S. 1748 — Kids Online Safety Act
Blackburn (R-TN), Blumenthal (D-CT), Thune (R-SD), Schumer (D-NY)
Ordered reported with a substitute · Aug. 5, 2026		 Introduced §103(a)(1)(D) gives minors options to opt out of personalized recommendations while keeping a chronological feed, and to limit types or categories of recommendations. Precedent for Step 2; minors only, no common rating. Compares introduced text; the substitute was not reviewed. Partial
H.R. 6253 — Algorithmic Transparency and Choice Act
Cammack (R-FL)
Subcommittee advanced · Dec. 11, 2025		 §2(a)(1)(D) requires an option for minors to limit the type or category of recommendations. Narrow precedent for Step 2; no rating standard or parental control of thresholds. Partial
H.R. 2657 — Sammy's Law
Wasserman Schultz (D-FL) + 5 bipartisan original cosponsors
Subcommittee advanced · Dec. 11, 2025		 §4 requires large platforms to give registered third-party safety software real-time interfaces to manage a child's content, interactions, and settings. Model for Step 4's third-party filters; no platform rating duty. Partial

State

Proposal or bill Relevant provisions and fit Fit
Louisiana — HB 427 / Act 552 (2026)
Signed May 29, 2026 · Effective Jan. 1, 2027		 Amends R.S. 9:2717.4(C): a duty of care with protective default settings for minors, parental access to a minor's connections, and optional parent notice when a minor is exposed to sexually explicit material. Parental-authority precedent for Step 3; notice after exposure, not filtering before delivery. Partial
California — SB 976 (2024)
Chapter 321
Feed provisions not enjoined · 9th Cir., Sept. 9, 2025		 §27002(b) lets verified parents require a nonpersonalized default feed, limit daily use, and hide like counts. Parental-control architecture for Step 3; governs how content is selected, not which categories are allowed. Partial
New York — A5114
In Assembly committee · Re-referred Jan. 7, 2026		 Proposed GBL §899-pp would require parental screen-time tools for child-targeted products, bar promotion of harmful or illegal activity, and let the state ban engagement features. Broader restriction approach; no family-set content ratings. Related

What this adds: The KIDS Act and KOSA would let minors limit "types or categories" of recommendations. This proposal names the five scales and their intensity levels, extends the controls to adults, puts parents in charge of minors' settings, and requires platforms to show that filtered content never arrives.

Notes

  1. Google Ads Help, "About content suitability," accessed September 2026. Inventory types apply to YouTube; "Moderate inventory" excludes "strong profanity, dramatized violence, and sexually suggestive themes," and "Limited inventory" also excludes moderate profanity. ↩

  2. YouTube Help, "Turn Restricted Mode on or off on YouTube," accessed September 2026: an "optional setting" that "works on the browser or device level, so you must turn it on for each browser you use." ↩

  3. Pew Research Center, "Teens, Social Media and AI Chatbots 2025," December 9, 2025. Self-reported; survey of 1,458 U.S. teens ages 13–17, September–October 2025. ↩

  4. Meta, "New 13+ Content Settings for Teen Accounts Expanding Globally on Instagram, Facebook, and Messenger," June 2026. Teens default to 13+; parents may choose a stricter "Limited Content" setting; teens need a parent's permission for "More Content." Meta notes it did not work with the MPA and that its systems are "not the same as a movie ratings board." ↩ ↩2

  5. Motion Picture Association, Classification and Rating Administration, "About Us," accessed September 2026: "In 2018, the Film Ratings System marked its 50th anniversary." ↩

  6. Entertainment Software Rating Board, "About ESRB," accessed September 2026: the industry's self-regulatory body, "Established in 1994." ↩

  7. 47 U.S.C. § 303(x) and statutory notes, including Telecommunications Act of 1996, Pub. L. 104-104, § 551(a)(9) (findings) and § 551(b)(1). The FCC accepted the industry's voluntary ratings in 1998, so § 551(b)(1)'s rating authority never took effect; the blocking feature was phased in for all covered sets by January 1, 2000. ↩ ↩2

  8. Meta Business Help Center, "About brand suitability controls and transparency tools," and "About inventory filter," accessed September 2026. Inventory filter settings are Expanded, Moderate, and Limited; topic exclusions (news, politics, gaming, religious and spiritual) apply to Facebook in-stream and Reels ads. ↩

  9. Cristos Goodrow, YouTube VP of Engineering, "On YouTube's recommendation system," YouTube Official Blog, September 15, 2021. ↩

  10. Pew Research Center, "Teens, Social Media and Mental Health," April 22, 2025. Survey of 1,391 U.S. teens ages 13–17, September–October 2024. ↩

  11. NetChoice v. Bonta, No. 23-2969 (9th Cir. Aug. 16, 2024), noting California could have used less restrictive means, "such as by (1) incentivizing companies to offer voluntary content filters or application blockers." A federal court in Colorado relied on the same passage in 2025. ↩

  12. H.R. 7757, KIDS Act, 119th Cong. § 214(b)(2)(A)–(B) (House-passed text, as referred to the Senate July 13, 2026); "teen" means ages 13 to 16 (§ 212). ↩

  13. H.R. 2657, Sammy's Law, 119th Cong. § 4 (introduced text). ↩