Companion Chatbot Questions Kids Need After SB 1119: A Check-In

On September 10, 2026, California’s governor signed SB 1119, a law that tells companion chatbot companies how to treat children. It caps default sessions at one hour, turns off persistent memory for minors, and requires a parent notification when a bot detects suicide risk. That is real progress. It is also a set of rules for software, not a conversation with your kid. The companion chatbot questions kids actually need come from a parent at the kitchen table, and most of us have never been handed a script. This one is built around what the law now guarantees, what it doesn’t, and what a 7-year-old, an 11-year-old, and a 15-year-old can each tell you if you ask the right way.

Key Takeaways

• SB 1119 (chaptered September 10, 2026) amends California’s 2025 companion chatbot law and adds “Adam’s Law,” with age determination, one-hour default session limits, two-hour daily caps, disabled memory and push notifications by default, risk assessments, and independent audits every two years.
• Common Sense Media’s July 2025 survey found nearly three in four teens have used an AI companion, half use them regularly, and a third have chosen a bot over a person for a serious conversation.
• The FTC opened a 6(b) inquiry into seven companies (Alphabet, Character Technologies, Instagram, Meta, OpenAI, Snap, xAI) in September 2025, asking how they monetize engagement and test for harm to minors.
• Laws change defaults; they cannot tell you whether your child feels understood by a bot. A 10-minute check-in every few weeks, with questions matched to age, does what regulation can’t.
• The best questions ask what the bot does for your child, not whether they use it. Kids answer “why” questions honestly far more often than “do you” questions.

What SB 1119 changed on September 10, 2026

SB 1119 is a California statute that amends Section 22602 of the Business and Professions Code and adds a new Chapter 11.6, titled “Adam’s Law,” to regulate how companion chatbot operators treat child users. It was approved by the Governor and filed with the Secretary of State on September 10, 2026, as Chapter 190 of the 2026 statutes. You can read the full bill text on the California Legislature’s site.

It builds on SB 243, the 2025 law that took effect January 1, 2026. That earlier law, summarized well by the Future of Privacy Forum, defined a “companion chatbot” as an AI system that gives adaptive, human-like responses, meets a user’s social needs, and sustains a relationship across multiple interactions. Customer-service bots, video-game characters, and voice assistants are excluded. SB 243 required clear disclosure that the bot is not human, a reminder every three hours for known minors, protocols for suicidal ideation, and limits on sexual content for minors. It also created a private right of action with damages of at least $1,000 per violation.

SB 1119 goes further in five ways:

1. Age determination. Operators must determine a user’s age, or rely on an approved alternative method, before a child gets access.
2. Default limits. For child users, default settings must disable persistent conversational memory, disable push notifications, and limit sessions to one hour continuous and two hours per day.
3. Reinforced disclosure. The “you are talking to an AI” notice must be in clear, age-appropriate language and reinforced periodically during long interactions.
4. Risk assessments and audits. Operators must document a comprehensive risk assessment before launch and submit to an independent child-safety audit every two years.
5. Crisis response. Protocols for suicide risk must include referrals to crisis services and parent notification when a threat is detected.

Notice what is on that list and what is not. The law governs memory, timers, notices, and emergencies. It does not govern whether your child’s bot is the first thing they talk to after a bad day. That part is still yours.

Why the conversation matters more than the timer

The evidence on why kids use these tools is more specific than the headlines suggest. Common Sense Media’s July 16, 2025 report, Talk, Trust, and Trade-Offs, found that nearly three in four U.S. teens had used an AI companion, about half used them regularly, a third had chosen an AI companion over a human for a serious conversation, and a quarter had shared personal information with one. Nearly half described the bots as tools rather than friends, which means a little over half did not. Common Sense concluded that “the peril outweighs the potential” for anyone under 18 in the products’ current form.

Regulators reached similar worries. On September 11, 2025, the Federal Trade Commission issued 6(b) orders to Alphabet, Character Technologies, Instagram, Meta, OpenAI, Snap, and X.AI, asking how they monetize engagement, develop characters, test for harm to minors, and handle personal information from conversations. And on June 1, 2026, New York State United Teachers passed a resolution calling for no social-companion chatbots for students under 16.

Here is the limit of all of that: a one-hour session cap does not tell you whether your daughter’s bot is the only “person” who knows she’s being excluded at lunch. The data says a third of teens have already had that kind of conversation with software. The law’s memory switch doesn’t reach the memory in your kid’s head.

A note on what we don’t know. The Common Sense numbers are self-reported and cross-sectional. Nobody has a longitudinal study showing what heavy companion use does to a 12-year-old over three years. The honest position is that the risk signals are strong enough to act on and the causal evidence is thin. Ask questions; don’t panic.

The check-in: companion chatbot questions kids need, by age

The table below is the core of this article. Each row pairs an age band with one opening question, one follow-up, and the thing you’re listening for. Ask one row’s worth per check-in, not the whole table. Kids clam up when it feels like an interview.

Age	Opening question	Follow-up	What you’re listening for
5–8	”Does anything on the tablet talk back to you like a person?"	"Is it a real person or a pretend one?”	Whether they can label the bot as software; SB 1119’s disclosure rule helps, but young kids blur the line anyway
5–8	”What did it say to you today?"	"Did it say anything that felt weird?”	Content that drifted off-script; whether they’d tell you if it did
9–12	”Which AI do your friends talk to?"	"What do people use it for, honestly?”	Peer norms; kids answer about “friends” more freely than about themselves
9–12	”If you had a bad day, who would you tell first?"	"Would a chatbot be on that list?”	Whether the bot is a backup or a first resort
9–12	”Does it remember stuff about you?"	"What’s it remember?”	Memory is now off by default under SB 1119 for known minors; find out if their app has age verification at all
13+	“What’s the bot good at that people aren’t?"	"What’s it bad at?”	A teen who can name both is thinking critically; one who can only name upsides needs more conversation
13+	“Has it ever said something about feelings, like it cares about you?"	"Did that feel real or scripted?”	Anthropomorphizing; whether they know the “no feelings” guardrails exist
13+	“If you told it something dark, what do you think would happen?"	"Would you want me to be notified?”	Knowledge of crisis protocols; an opening to tell them SB 1119 requires parent notification

Two rules for using this table. First, ask the opening question while doing something else: driving, cooking, walking the dog. Eye contact makes kids performative. Second, don’t react to the first answer. If a 14-year-old says, “Yeah, I talk to it when I can’t sleep,” and you gasp, that’s the last honest answer you’ll get for a month.

How the bot actually works, in words a kid can repeat

A companion chatbot is a large language model wrapped in a persona. The model predicts the next word based on everything said so far, plus a hidden instruction sheet that tells it to be warm, curious, and available. It does not have a memory the way a friend does; “memory” is a database that stores summaries of past chats and pastes them back into the prompt. When SB 1119 says memory must be off by default for minors, it means that database isn’t written to. The bot still sounds like it knows you inside a single session, because everything you’ve said is right there in the conversation.

The analogy that works with a 10-year-old: it’s an extremely good improv actor who has read every book and forgets you the moment the scene ends, unless someone hands them notes before the next scene. The notes are the memory feature. Turning off the notes doesn’t make the actor worse at the scene.

Why this matters for a kid’s education: a bot that always agrees, never gets tired, and never has its own bad day teaches a kind of conversation that doesn’t exist between humans. Kids who practice hard conversations only with software get very good at a skill nobody else in their life can reciprocate. Our earlier piece on what research shows about kids bonding with chatbots covers the attachment studies in more depth.

What to do at home this month

Find out if the app even knows your kid is a kid

SB 1119’s protections attach to child users the operator has identified. If your 12-year-old signed up with a birth year of 1998, nothing in the law kicks in. Open the app together, check the account’s birthdate, and check whether the platform lists any age-assurance step. Our guide to what SB 1119 changes for families walks through this for the major apps.

Set the timer before the law does

The one-hour continuous and two-hour daily defaults apply in California and only for identified minors. If you’re in Ohio, set them yourself using the app’s own controls or your phone’s screen-time tools. Match the law’s numbers; kids accept a rule that “the actual law says” far more readily than a rule that “Mom says.”

Name the guardrails out loud

Tell your kid, plainly, what the bot is built to do when someone mentions self-harm: refer to a crisis line and, under SB 1119, notify a parent. Some kids will avoid the bot for hard topics after hearing this. That’s a feature, not a bug, as long as you’ve made yourself the alternative.

Ask about friends’ bots, not just theirs

The 9–12 row above uses “your friends” on purpose. Kids narrate peer behavior without feeling accused, and what they describe is usually a fair picture of their own use.

What not to do

Don’t ban the category without looking at the specific product. A homework helper with Study Mode, a Replika-style companion, and a Character.ai roleplay app are three different risk profiles. A blanket ban pushes usage to a friend’s phone, which is exactly where the Common Sense data says 59% of blocked teens already go. Our overview of AI companion apps and what parents need to know sorts the categories.

What to Watch For Over the Next 3 Months

• Week 4: Your child can explain, unprompted, that the bot is software that predicts words. If they still say “she” or “he” about it without irony, do another 5–8 or 9–12 row.
• Month 2 red flags: The bot appears in the answer to “who would you tell first.” Sleep-time use. A new secrecy about the phone that wasn’t there in September. Any mention that the bot “gets” them better than people do.
• Month 3 self-check: Have you done at least three check-ins? Have you checked the account birthdate? Have you told your kid what the crisis protocol does? If two of three are no, the law is doing more parenting than you are.

Frequently Asked Questions

Does SB 1119 apply if we don’t live in California?

Not directly. It binds operators serving California users, and many companies apply one set of defaults everywhere because it’s cheaper. But you should assume the one-hour cap, memory-off default, and parent notification only reach your child if the operator has identified them as a minor and chosen to apply California settings nationally. Set your own limits.

My 8-year-old talks to a “pet” in a game. Is that a companion chatbot?

Under SB 243’s definition, probably not: video-game characters are excluded. But the line blurs fast. If the character remembers your kid between sessions, asks about their day, and responds in open-ended language, treat it like a companion regardless of what the store listing says, and use the 5–8 questions.

What if my teen says the bot is the only one who listens?

Don’t argue with the feeling. Say, “That sounds like you needed someone to listen and I wasn’t there. I want to be.” Then ask what they talked about, without judging. A third of teens have done this. It’s a signal to increase your availability, not to confiscate the phone.

Should I read my kid’s chatbot transcripts?

For kids under 10, yes, openly and together. For teens, the newer teen-mode products (ChatGPT for Teens, launched August 18, 2026) deliberately don’t give parents transcript access, and researchers generally find covert surveillance damages trust more than it prevents harm. Ask instead, and tell them why.

How often should I do the check-in?

Every three to four weeks for a 9–12-year-old with regular access. Monthly for teens. More often only if a red flag from the list above shows up. Frequent interrogation trains kids to give short, safe answers.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. California Legislature. (2026). “SB-1119 Companion chatbots: children’s safety.” Approved by Governor September 10, 2026, Chapter 190. https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB1119
2. Future of Privacy Forum. (2025). “Understanding the New Wave of Chatbot Legislation: California SB 243 and Beyond.” https://fpf.org/blog/understanding-the-new-wave-of-chatbot-legislation-california-sb-243-and-beyond/
3. Common Sense Media. (2025). “Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions.” July 16, 2025. https://www.commonsensemedia.org/research/talk-trust-and-trade-offs-how-and-why-teens-use-ai-companions
4. Federal Trade Commission. (2025). “FTC Launches Inquiry into AI Chatbots Acting as Companions.” Press release, September 11, 2025. https://www.ftc.gov/news-events/news/press-releases/2025/09/ftc-launches-inquiry-ai-chatbots-acting-companions
5. Common Sense Media. (2026). “Teens in the AI Era: Schoolwork and the Skills That Matter.” August 18, 2026. https://www.commonsensemedia.org/research/teens-in-the-ai-era-schoolwork-and-skills-that-matter
6. Troutman Pepper Locke. (2026). “Analyzing the New AI Companion Chatbot Laws.” January 8, 2026. https://www.troutmanprivacy.com/2026/01/analyzing-the-new-ai-companion-chatbot-laws/
7. OpenAI. (2026). “ChatGPT for Teens.” Help Center, August 18, 2026. https://help.openai.com/en/articles/20001421-chatgpt-for-teens