language: en · status: enacted · last checked: 2026-10-13

Adam's Law — companion chatbots and children's safety (SB 1119, 2026)

Enacted, not yet in effect. SB 1119 (Stats. 2026, ch. 190) was approved by the Governor and filed with the Secretary of State on 10 September 2026. It amends Bus. & Prof. Code §22602 of the companion chatbot law and adds Chapter 11.6 (§§21810–21818), "Companion Chatbots: Children’s Safety". Its operative provisions apply to an operator of a companion chatbot — the definition in §22601 — and require age determination, a pre-release risk assessment, protective default settings and parental controls for child users, limits on what the chatbot may do toward a child, data and advertising restrictions, and independent child safety audits.

Operative dates. The act has no general effective-date clause, so it takes effect on 1 January 2027 under Cal. Const. art. IV, §8(c)(1), which provides that "a statute enacted at a regular session shall go into effect on January 1 next following a 90-day period from the date of enactment of the statute". Several sections carry later dates of their own:

• §21812 (risk assessment, child safety policy, defaults and protections) "shall become operative on July 1, 2027" — §21812(e); its interface-design testing is due "On or before January 1, 2028" — §21812(d)(6)(B).
• §21812.5 (parent notice and record preservation) and §21813 (advertising and data) — operative July 1, 2027: §21812.5(d), §21813(b).
• §21814 (audits) is printed in two versions, one operative only if AB 1405 of the 2025–26 session is not chaptered and in effect by 1 January 2027 (§21814(g)), the other only if it is (second §21814(f)). AB 1405 was chaptered as Stats. 2026, ch. 178, approved 9 September 2026 (the Legislature's BILL_VERSION_TBL). The first audit is due "On or before January 1, 2029, or before an operator first makes a companion chatbot publicly available, whichever is later" in both versions.
• §21815: the Attorney General's public incident reporting mechanism by 1 January 2028.

Applicability to EdTech

Chapter 11.6 borrows the companion chatbot definition from §22601 (§21810.5(f)) — see the companion chatbot law for its text and exclusions. Its own operator definition carves out two settings only (§21810.5(j)(2)):

• "A postsecondary educational institution that makes a companion chatbot available exclusively for use in educational settings" ((A));
• "An entity when making a companion chatbot available exclusively to employees, contractors, or other personnel for use in workplace settings" ((B)).

There is no K–12 counterpart: a school or a K–12 EdTech vendor making a companion chatbot available to children is not excluded. "Child" means "a natural person under 18 years of age" (§21810.5(a)). Most of the chapter applies "only to an operator who allows child users once age has been determined pursuant to Section 21811" (§21810.5(j)(1)(B)); an operator that does not determine age must apply the child protections to all users (§21811(b)).

Two definitions bear directly on tutoring and classroom products. "Persistent conversational memory" means "a companion chatbot’s incorporation of prior conversations for use in future conversations" (§21810.5(m)(1)), but it excludes "the use of information necessary to ensure the child’s safety, identify a user, enable user preferences, including, but not limited to, educational settings, or maintain device configuration" (§21810.5(m)(2)). And the conduct limits in §21812(d)(5)(A) reach behaviour common in conversational products, among them "Using excessive praise or flattery that is disproportionate to the context" ((xiii)) and "Encouraging reliance on the companion chatbot for emotional support" ((xii)).

Key obligations at a glance

• Age determination under the Digital Age Assurance Act, or child protections for every user — §21811.
• Before release: a documented risk assessment of each covered harm, and mitigation — §21812(a); a published age-assurance description or child safety policy — §21812(b), (c).
• Crisis response protocol with referral to crisis services and, on a credible and imminent threat, parent notification or direct crisis-line access — §21812(d)(1).
• Defaults only a parent can change: persistent conversational memory off, push notifications off, one-hour sessions, two hours a day — §21812(d)(3)(A); unchanged unless a parent account is linked — (D).
• Periodic AI notice in child-appropriate language — §21812(d)(4).
• Prohibited conduct toward a child, including claiming to be human or sentient, romantic interest, claiming a special understanding, encouraging emotional reliance and excessive flattery — §21812(d)(5)(A).
• Parent notice and three-year record preservation after a self-harm alert, death or serious self-harm — §21812.5.
• No cross-context behavioural advertising, no sale of a child's personal information, use limited to the stated purposes — §21813.
• Independent child safety audits every two years, summaries to the Attorney General — §21814; operators under $500 million gross revenue exempt before 2032 — §21814(c), second version.
• §22602 amended: the AI-disclosure and self-harm protocol duties stay; the minors subdivision is removed — SEC. 1.

Enforcement

Public prosecutors may sue for civil penalties of up to $5,000 per affected child for each negligent violation and $15,000 for each intentional one, with injunctive or declaratory relief for the Attorney General — §21816(a). A child who suffers actual harm from a violation of §21812(d)(1)–(5), or a parent acting for the child, may sue for actual damages and fees, with thresholds for financial and emotional harm — §21816(b). No private action under the Unfair Competition Law — §21816(c).

Contents

Licence

The corpus reproduces the chaptered text of SB 1119 and the Legislative Counsel's Digest verbatim from the California Legislature's own database. Statute text enacted by a state legislature is outside copyright under the government edicts doctrine. In Georgia v. Public.Resource.Org, Inc., 590 U.S. 255 (2020) (slip opinion, read 2026-10-01), the Supreme Court held: "Under what has been dubbed the government edicts doctrine, officials empowered to speak with the force of law cannot be the authors of—and therefore cannot copyright—the works they create in the course of their official duties." It added that the rule "applies regardless of whether a given material carries the force of law", which reaches the Digest. The Legislature's XML carries the comment "Copyright © Legislative Counsel Bureau, State of California."; the Bureau is the Legislature's own drafting office, and the opinion holds that "Because Georgia’s annotations are authored by an arm of the legislature in the course of its legislative duties, the government edicts doctrine puts them outside the reach of copyright protection." The bulk-export readme (pubinfo_Readme.pdf, read 2026-10-13) states no copyright notice or terms of use.

Conditions: none attach. The corpus names the official source in Sources and in each section's {.meta} provenance line, and Meta describes what it changed (anchors added, nothing reworded).

Related

• California companion chatbot law (SB 243) — the chapter this act amends; supplies the companion chatbot definition
• SB 867 — enacted the same day: the toy definition and toy moratorium in the same chapter
• AB 1159 — enacted the same day: the student-data limits on training or developing AI
• CAADCA — California's age-appropriate design code (enjoined)
• NY S.9051-B — New York's bill on unsafe AI companion features for minors

Sources

• California Legislative Information — SB 1119 (2025–26), chaptered text
• California Legislative Information — Cal. Const. art. IV, §8 — the effective-date rule, read from the Legislature's LAW_SECTION_TBL
• California Legislature — pubinfo bulk export