Harm Disclosure + Warning Label

Instagram's feed predicts whether the first post you see will "keep you spending time in your Feed."¹ That goal is spelled out on a company website, not in law, and what companies know about harm reaches parents secondhand, from the press. Congress should require large platforms to state what their feeds optimize for, share their evidence of harm, and show a plain warning at the start of every session.

The Problem

Thirty-six percent of American teens say they are on at least one major platform almost constantly.² The feeds are engineered to hold that attention. YouTube added watch time to its recommendations in 2012 and accepted an immediate 20% drop in views to do it; one of its engineering vice presidents has written of staying up late on random videos "when I could have instead been learning a new language."³

Ranked feeds pursue measurable goals: minutes watched, return visits, clicks, or the company's own estimate of satisfaction. YouTube separates raw watch time from "valued watchtime," the viewing people rate as worthwhile in surveys.³ Those are business choices, and they decide what a person sees next.

Minnesota, California, and New York have enacted mental health warnings, but none of their labels says what the feed is for. Three gaps keep families guessing:

1. The goal stays hidden. Instagram's ranking models estimate, among other things, "how much time you're likely to spend on this post."¹ Meta's own explainer lists the predictions but not how much each one counts.
2. The evidence stays inside. In 2019, Instagram's researchers found that one in three teen girls who struggled with body image said Instagram made them feel worse; the public learned of it in 2021, from The Wall Street Journal.⁴
3. The warning stops at state lines. Minnesota's label took effect July 1, 2026.⁵ In January 2027, California's will say social media "has not been proven safe for young users,"⁶ while New York's will say excessive use "may be associated" with significant harm.⁷

Why legislation: In 2024 the Surgeon General asked Congress to require a warning label, citing tobacco studies showing that labels "can increase awareness and change behavior."⁸ In April 2026, the Senate Commerce Committee advanced a Britt–Fetterman bill to do so.⁹ Federal law has required a warning on every cigarette pack since 1966.¹⁰ The FTC can punish a misleading claim about a feed, but no federal law requires an honest one.¹¹ Companies that profit from Americans' attention should say plainly what they do with it.

The Solution

A three-step staircase: each step stands alone, and each step up brings the company's knowledge closer to the moment of use. Scope: social media and video platforms above 10 million monthly U.S. users, for users of all ages.

Step 1 — Say what the feed is for. Platforms publish, in plain language, the main goals each recommender optimizes (time spent, return visits, ad revenue, survey-measured satisfaction) and their relative weight, and update the statement when the goals change. KOSA's introduced text would require a similar disclosure of the "quantities" an algorithm is designed to optimize,¹² while exempting trade secrets. People cannot weigh a bargain whose terms they never see.

Step 2 — Hand over the evidence. Platforms give the FTC their internal studies and serious-harm trends, then publish summaries that separate established findings from associations and open questions, including null and beneficial results. Parents should not have to learn a company's findings from a newspaper.

Step 3 — Warn at the door. At the start of each app session, show a brief notice stating the feed's main goal and the risks the evidence supports, with one-tap links to time limits, a nonpersonalized feed, and notification controls. The FTC, with HHS, writes the wording, tests it for comprehension and fatigue, and may approve an equally effective format. A national label gives every family the same facts.

Where to start: Step 1 is the floor; it asks companies only to state facts about their own products. Step 3 is the heart of the proposal.

Administration and enforcement: The FTC, consulting HHS and independent researchers, issues rules within 12 months; compliance follows a year later. Civil penalties apply to missing or false disclosures and obstructed audits, and the FTC revisits the label as evidence changes.

Risks and Mitigations

• Compelled speech: This is the main risk. In 2025 a federal judge blocked Colorado's warning pop-ups as likely compelled noncommercial speech that fails strict scrutiny,¹³ and NetChoice is suing to stop Minnesota's every-session label.¹⁴ Steps 1 and 2 ask only for facts about a company's own product and research, which stand on firmer ground than Step 3's warning. The risk remains.
• Warnings people stop seeing: The FDA found that cigarette warnings unchanged for 35 years "go largely unnoticed and unconsidered."¹⁵ Hence short, tested, periodically refreshed wording, paired with links that let people act. A label alone won't change a habit.
• False certainty: Researchers dispute how large the harms are. The label claims only what the evidence supports, in the Surgeon General's own terms ("associated with"),⁸ and changes as findings do.

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

State

What this adds: State labels warn of mental health risks in words a legislature chooses. This proposal adds what none of them requires: a plain statement of what each feed is built to maximize and the company's own evidence of harm, delivered nationwide at the moment people open the app.