The New York AI chatbot law signed on October 28 bars companion-style chatbots from acting like a friend, partner, or real person to anyone under 18. It starts on January 1, 2027, and companies that break it face fines of up to $25,000 per violation.
Here are the facts worth knowing before the rules kick in.
- Signed by Gov. Kathy Hochul on October 28, 2026 (bill S.9051-B/A.10379-C).
- Takes effect January 1, 2027, according to IAPP.
- Covers chatbot providers and developers of companion chatbots.
- Penalties run up to $25,000 per violation, enforced by the state attorney general.
If you build, run, or just use these apps, the next few months will bring visible changes, mostly in how apps check your age.
What the New York AI chatbot law actually bans
The law targets designs that make a bot feel like a person in a child’s life. Reporting from ID Tech lists the restricted features, and the table below groups them.
| Restricted behavior | What it means in practice |
|---|---|
| Posing as a human | A bot cannot claim to be a real person, be alive, or have human emotions. |
| Claiming a relationship | No pretending to be a friend, partner, professional, or authority figure to a child. |
| Using private details | No unprompted emotional questions and no reuse of a child’s health or personal information from past chats. |
| Harmful encouragement | No promoting or assisting suicide, self-harm, eating disorders, or drug and alcohol use. |
| Isolation tactics | No urging kids to keep secrets or to avoid parents, trusted adults, or professionals. |
| Sexual content | No sexually explicit content aimed at minors. |
Hochul said minors will lose access to chatbots that pretend to be their friends, boyfriends, or girlfriends. That is the core of the change.
How age checks will work
Companies cannot offer the restricted features to anyone until they confirm the user is not a minor. That makes age assurance the part most people will actually notice.
Providers must offer more than one way to verify age, and at least one cannot rely only on a government ID. They must also delete the age data once verification is done. If you have already been through an age check on Discord, expect something similar.
The deletion rule matters because the biggest worry with age checks is a pile of sensitive data sitting on a server. Here it is meant to be wiped right after use.
What happens to companies that break the rules
The attorney general can seek civil penalties of up to $25,000 per violation. Other remedies include damages, restitution, and orders to destroy unlawfully collected data.
ID Tech also notes the destruction can reach AI models trained on that data. That is a heavy penalty for a company whose product depends on its model.
New Yorkers will also get a public website run by the attorney general to file complaints about unsafe chatbot behavior. Attorney General Letitia James said the law “sets clear legal parameters to prevent that harm and keep kids safe.”
What parents and teens will notice first
Most families will see this as new sign-up steps, not new bans. Apps that serve both adults and teens may ask for proof of age before unlocking companion features.
Some bots may simply feel more formal for unverified users. A chatbot that stops saying “I missed you” is the law working as intended.
The rules sit alongside a wider wave of pressure on AI companies over kids. Our coverage of the Texas lawsuit against OpenAI and the “unacceptable risk” rating for ChatGPT for Teens shows how fast this is moving.
The bill was sponsored by Sen. Kristen Gonzalez. According to ID Tech, Common Sense Media worked on it with the governor, the sponsors, and the attorney general, which suggests the child-safety lobby will watch closely how it is applied.
For app makers, the practical lesson is that “companion mode” can no longer be the default for everyone. Features like persistent memory of a child’s feelings or health details need to be switched off until age is confirmed. That is a real product change, not a legal footnote.
Companies serving users outside New York may simply apply the same rules everywhere, because maintaining two versions of a chatbot costs more than one stricter version. Whether that happens is unconfirmed, so treat it as a likely trend and not a promise.
Frequently Asked Questions
When does the New York AI chatbot law start?
The law takes effect on January 1, 2027, per IAPP. One other report only says “in the coming months,” so check dates with your provider.
Does it ban AI chatbots for kids entirely?
No. It bans specific unsafe features, such as posing as a human or a romantic partner. Regular homework or search help is not what it targets.
Do adults need to prove their age?
Possibly. Providers cannot offer the restricted features until they confirm a user is not a minor, so some adults may see an age check.
What is the penalty for breaking it?
Up to $25,000 per violation, plus possible damages, restitution, and orders to delete unlawfully collected data.
Our take: right goal, tough to enforce
Stopping bots from posing as a child’s best friend is a reasonable line, and the data-deletion rule is a smart touch. The hard part will be enforcement, since a determined teen can still find a way around an age gate. If you run an app with a chat feature, start auditing now. If you are a parent, talk to your kids about what these bots are and are not.
Leave a Reply