$1 Million per Child: California Raises the Cost of Unsafe Feeds

California Governor Gavin Newsom signed a package of child online-safety laws on September 10, 2026. The state’s package announcement confirms separate measures restricting addictive platform features for users under 16, strengthening companion-chatbot safeguards and extending protections for children’s data.

The headline’s $1 million figure is a conditional liability ceiling, not an automatic fine for showing a child an algorithmic feed. Associated Press coverage of the September 10 signing confirms that qualifying social-media companies may face up to $1 million per child when negligence on their platforms causes harm; feed restrictions, chatbot duties and school-device rules come from different bills.

The package creates five separate compliance tracks

The measures apply to different operators, age groups and conduct. Their principal requirements should be read separately:

• Addictive features — AB 1709: Covered platforms must keep addictive feeds, autoplay and any subsequently designated addictive feature away from California users under 16. Age verification and public enforcement support that prohibition.
• Child-injury liability — AB 2: Social-media platforms with more than $100 million in annual gross revenue face enhanced statutory damages when a failure to exercise ordinary care causes injury to a child under 18.
• Companion chatbots — SB 1119: Operators must assess child-safety risks before making a new or substantially modified chatbot available in California. Operators that permit child users must add protective defaults, parental controls, crisis procedures and limits on specified manipulative conduct.
• Chatbot audits and reporting: Covered operators must arrange independent child-safety audits, submit summaries to the attorney general, publish high-level findings and provide a way for third parties to report safety incidents.
• School devices — SB 1128: Beginning with the 2027–28 school year, local educational agencies generally cannot require kindergarten pupils, including transitional-kindergarten pupils, to take school-issued electronic devices home. This is not a statewide device opt-out for every student.

AB 1709 restricts features rather than every account

AB 1709 covers websites, online services and applications that make an addictive feature a significant part of their service. Its definition includes personalized feeds that recommend user-generated media using information associated with a user or device, as well as autoplay. Services limited to commercial transactions or consumer reviews, and feeds used primarily for cloud storage, are excluded.

The law does not categorically prohibit an under-16 user from creating or maintaining an account. A platform may keep the account available if it does not provide prohibited features. Before offering an addictive feature, however, the platform must verify the user’s age through California’s age-assurance framework or use the alternative age determination specified by state law.

Only the attorney general or a local public prosecutor may bring an AB 1709 enforcement action. A court may impose up to $25,000 per affected minor for a negligent violation or $50,000 for a knowing violation, taking account of the platform’s size, the violation’s severity and duration, and good-faith compliance efforts. Those sums are public civil penalties, not payments a family automatically receives.

The $1 million ceiling requires negligence and injury

AB 2 addresses a broader category of platform harm than AB 1709’s feature prohibition. It applies when a covered social-media platform violates its ordinary-care duty by causing injury to a child under 18. The defendant must also generate more than $100 million a year in gross revenue.

When those conditions are met, statutory damages are the larger of $5,000 per violation, capped at $1 million per child, or three times the child’s actual damages. An addictive-feed violation alone therefore does not establish a $1 million award: a claimant still needs injury, a failure of ordinary care and a platform meeting the revenue threshold.

The remedy is cumulative rather than exclusive, but it is prospective. AB 2 does not apply to a legal case pending on or before January 1, 2027, and its enhanced-damages provision is scheduled to expire on January 1, 2035 unless lawmakers extend or replace it.

Adam’s Law sets a different regime for companion chatbots

SB 1119, known as Adam’s Law, regulates people or companies that make companion chatbots available to California users. Before releasing a new or substantially modified chatbot, an operator must document foreseeable risks of physical, financial, severe psychological or emotional harm, serious privacy intrusion and unlawful discrimination, then document reasonable mitigation measures.

For operators that allow children to use their chatbots, SB 1119’s enacted requirements include crisis referrals, age-appropriate AI notices, parental controls, disabled push notifications by default, a one-hour continuous-session limit and a two-hour daily limit. Parents may adjust those defaults and can disable access for a child under 16; without a linked parent account, the defaults cannot be changed.

Operators must also take reasonable measures to prevent a child-facing chatbot from encouraging self-harm, claiming to be human or conscious, simulating romantic interest, discouraging breaks, fostering emotional dependence or urging a child to conceal usage. The law separately restricts behavioral advertising, sales of information gathered through child chatbot conversations and dark patterns affecting safety controls.

Most of these operating duties become operative July 1, 2027. Public prosecutors may seek up to $5,000 per affected child for each negligent violation and $15,000 for each intentional violation. A child who suffers actual harm from specified failures—or a parent acting for the child—may seek actual damages and other relief, but not every technical violation creates a private claim.

Audits and school-device rights follow their own schedules

Initial companion-chatbot audits are due by January 1, 2029, or before an operator first makes the chatbot publicly available, whichever is later. Audits generally recur every two years, with an earlier audit required before a substantial modification when the relevant risk assessment shows increased child-safety risk. Before 2032, operators with less than $500 million in prior-year gross revenue are exempt from the audit section.

An operator must submit an audit summary to the attorney general within 30 business days of receiving the report and publish a high-level summary within 90 days. California’s attorney general must establish a separate public complaint mechanism for companion chatbots by January 1, 2028.

SB 1128 is narrower than a general family right to reject school technology. From the 2027–28 school year, a district, county office of education or charter school may authorize a kindergarten or transitional-kindergarten pupil to take a device home but generally may not require it. Exceptions cover a device needed under an individualized education program, during an emergency that prevents in-person attendance or for an independent-study program.

The signed package therefore raises several distinct forms of accountability: feature-specific public penalties, conditional child-injury damages, chatbot controls and audits, and a limited rule for devices sent home with the youngest pupils. The $1 million ceiling belongs only to AB 2’s negligence-and-injury framework, even when the alleged harm involves the kinds of engagement systems addressed separately by AB 1709.

Also read:

• Meta and YouTube Appeal $6 Million Social Media Addiction Verdict
• The Death of a VR Unicorn: Rec Room Shuts Down After 150 Million Users, $294 Million Raised, and a $3.5 Billion Valuation
• Norman Angell Was Misread: Trade Raises War’s Cost but Cannot Guarantee Peace