New York SAFE for Kids Act final rules (2026)

New York — SAFE for Kids Act Final Rules: Quantified Age-Assurance Accuracy Minimums for Addictive Feeds (Rules July 28, 2026; Act Effective January 25, 2027)
New York · N.Y. Off. of the Att'y Gen., SAFE for Kids Act Final Rules (July 28, 2026) (Act eff. Jan. 25, 2027)

Status
Enacted (not yet in effect)
Takes effect
January 25, 2027
Applies to
Social media companies covered by the SAFE for Kids Act operating in New York, which must select a qualifying age assurance method, test it annually and retain results for five years; age assurance vendors supplying those methods.
Enforced by
New York State Office of the Attorney General

New York's Stop Addictive Feeds Exploitation (SAFE) for Kids Act got its final implementing rules on July 28, 2026, released jointly by Attorney General Letitia James and Governor Kathy Hochul. The Act itself takes effect January 25, 2027. What makes these rules significant is that they put numbers on age assurance where almost every other regime says only that a method must be 'reasonable.' The rules set an Accuracy Minimum: a method's false positive rate must be no more than 0.1% for minors aged 0 to 7, 1% for ages 8 to 13, 2% for ages 14 to 15, 8% for age 16 and 15% for age 17, excluding cases where a user refuses or fails to provide data and cases where the result is inconclusive. The method must also detect attempts to circumvent it at a rate of 98% or better. A second standard, the Total Accuracy Minimum, applies the same percentage ceilings to false positives and inconclusive outcomes combined — closing the obvious gap, since a method could otherwise hit its accuracy target by returning 'inconclusive' whenever it was unsure. Social media companies must pick a method meeting these rates, test it annually, and keep the test results for at least five years. The graduated thresholds are the design insight: being wrong about a seven-year-old is treated as far more serious than being wrong about a seventeen-year-old.

Who is protected: New York minors, especially the youngest age bands where the permitted false-positive rate is most stringent — protected from addictive social media feeds by age assurance that must meet quantified, annually tested accuracy floors rather than an undefined reasonableness standard.

Legal detail

New York Office of the Attorney General final rules implementing the Stop Addictive Feeds Exploitation (SAFE) for Kids Act, released July 28, 2026; the Act takes effect January 25, 2027. QUANTIFIED ACCURACY STANDARDS, which is what distinguishes this from 'reasonable age assurance' regimes elsewhere. Accuracy Minimum — maximum false positive rate by age band: 0.1% (ages 0-7), 1% (8-13), 2% (14-15), 8% (16), 15% (17), computed excluding user failures or refusals to provide requested data and excluding inconclusive outcomes; plus a circumvention-detection rate of at least 98%. Total Accuracy Minimum — the same age-banded ceilings applied to false positives AND inconclusive outcomes COMBINED, plus the same 98% circumvention-detection floor. The second standard forecloses the gaming route the first would otherwise leave open, where a method meets its false-positive ceiling by classifying uncertain cases as inconclusive rather than wrong. Compliance obligations on covered social media companies: select an age assurance method meeting the accuracy rates, conduct annual testing, and retain testing results for a minimum of five years. Regulatory significance: the graduated thresholds scale tolerance for error inversely with the vulnerability of the age band, and the retention requirement makes the testing auditable after the fact rather than self-attested at a point in time. Final rule published by the Attorney General at ag.ny.gov/safe-for-kids-act-final-rule.

Key facts

JurisdictionNew York
LevelState
StrengthStronger
EnactedJuly 28, 2026
CitationN.Y. Off. of the Att'y Gen., SAFE for Kids Act Final Rules (July 28, 2026) (Act eff. Jan. 25, 2027)
Private right of actionNo, agency enforcement only
PenaltiesEnforcement by the Attorney General under the SAFE for Kids Act; specific penalty provisions not captured from the rules summary
Topicschildren's online safety · consumer data privacy · biometric data · automated decision-making · AI disclosure and transparency
Agency pageSAFE for Kids Act Final Rule — New York State Attorney General

More AI rules in New York

Related children's online safety rules elsewhere

See something wrong or out of date? Submit a correction. Every entry must carry a verifiable source.