What Are the Key Dates?
| Date | Law | What happens |
|---|---|---|
| November 5, 2025 | New York, AI Companion Models Law | First state companion-chatbot law takes effect |
| January 1, 2026 | California SB 243 | Companion chatbot law takes effect |
| October 28, 2026 | New York S9051B / A10379C | Governor Hochul signs the new law |
| **January 1, 2027** | New York S9051B / A10379C | "Unsafe AI companion features" barred for minors and unconfirmed users |
| **July 1, 2027** | California SB 1119, "Adam's Law" | Main requirements take effect, per FIRE |
What Does the New York Law Require?
According to Holland & Knight and Blank Rome:
- It adds a new Article 48 to the General Business Law and, from January 1, 2027, prohibits operators from offering "unsafe AI companion features" to minors in New York.
- The statute lists 12 categories of unsafe features, including outputs that suggest the chatbot is human, flatter the user, or encourage self-harm or secrecy.
- Operators may not use a user's health or personal information acquired more than 12 hours earlier or in a prior session.
- Operators must offer more than one age-assurance method, and age data may be used only to determine age and must be deleted right after.
A commentary from Communications Law Group adds that the covered definition is broad and could reach general-purpose chatbots as well as products marketed as companions. For the signing itself, see Fortune's report and our New York explainer.
What Does California's Adam's Law Require?
Per FIRE's analysis, operators can choose between relying on a standardized age-range indicator under California's Digital Age Assurance Act, under which operating systems collect a user's age at account setup, or extending the law's child restrictions to everyone. FIRE warns that age assurance can involve identity verification, especially where methods include a government ID, and that anonymous access to these conversations will become harder.
What Will Age Checks Look Like?
The coverage doesn't give one method. New York allows methods permitted under its existing age-assurance rules, and operators must offer more than one. California points toward a device-level age signal. Either way, expect a prompt or a signal, not just a tick-box, on platforms serving those states.
What Does AI Companion Age Verification Mean for Adult Users?
1. More checks. If a platform can't confirm your age, the New York law treats you like a minor for the restricted features. 2. Less anonymity. Age assurance may involve ID or a device signal. 3. Privacy of age data matters. New York requires deletion right after the check, which is worth confirming a platform actually does. 4. Check each platform's notices for changes to terms or availability.
See do AI girlfriend apps sell your data and is an AI girlfriend safe in 2026.
What Don't We Know?
How each platform will respond. None on our roster has announced changes.
Which methods platforms will choose. The statutes allow several.
How broadly "covered" will be read. Commentators say it could be wide.
We didn't read the bill text, so confirm details there. An earlier note of ours said New York's effective date wasn't in the coverage we had. Law-firm summaries now give January 1, 2027. One report also puts state AI bills enacted in 2026 at 280.
10 AI Companion Apps With Stronger Privacy Practices in 2027
End-to-end encryption and a written no-training-on-chats policy, the strongest privacy posture on our roster. See the Clonella review.
Free with no credit card, so there is no payment record. Unlimited chat with no message cap or card, over 1,000 companions and face-consistent images. Memory is basic on free and there is no voice. See the Selira AI review.
Operated from Switzerland, with deep memory on its top tier. It is web-only, has about 18 pre-made characters, and coins meter images and video. See the Darlink AI review.
Asks for no ID, legal name, phone or address and confirms age by your own declaration, an approach that may change if age-assurance rules apply to it. Fixed-length passes with unlimited chat, locked faces and a light-touch privacy design. See the Lovora AI review.
Published Estonian corporate registration and discreet billing. See the AI Angels review.
States that no adult transaction appears on your bank statement. Voice calls and a media allowance on its own models, with discreet billing stated on its pricing page. New, with no independent reviews yet. See the Sparkles AI review.
Reportedly crypto only, so no card appears on the transaction. In-house characters, an editable memory panel and group rooms of up to four. See the Erolove AI review.
Accepts crypto and Telegram Stars alongside cards. Four image engines, video with its own audio and unlimited text chat on paid tiers. Coins never expire. Browser-only. See the AI Peeps review.
A named Delaware and UK-linked operator with no breach found, but no end-to-end encryption. A deep character builder with 10 free image generations and 15 free messages a day. Credits meter the extras, and plans changed during 2026. See the Nectar AI review.
Run by EverAI Limited in Malta and billed as Everai, though tokens can raise the real cost. Native iOS and Android apps, polished images, an Anime tab and 120-second video, with a token system that can raise the real cost. See the Candy AI review.
Also Worth Exploring
FAQs
When does New York's AI companion chatbot law take effect?
January 1, 2027, according to Holland & Knight. Governor Hochul signed it on October 28, 2026.
When does California's Adam's Law take effect?
Its main requirements take effect July 1, 2027, per FIRE, though an earlier summary gave January 1, 2027.
Will AI girlfriend apps ask for ID?
Possibly more age verification on platforms serving those states. The laws allow several methods, and no platform on our roster has announced changes.
Does the law ban adults from AI companions?
No. It restricts specified features for minors and for users whose age isn't confirmed, and requires operators to use age assurance.
Is my age data kept?
New York requires age-assurance data to be used only to determine age and deleted right after.
What are the unsafe AI companion features?
Twelve categories, including suggesting the chatbot is human, flattery and sycophancy, and encouraging self-harm or secrecy. If you or someone you know is struggling, support is available. In the US you can call or text 988 to reach the Suicide and Crisis Lifeline.
Sources
- Holland & Knight, "New York Enacts AI Chatbot Safeguards for Minors", November 5, 2026.
- Blank Rome, "The BR Privacy, Security & AI Download: November 2026".
- Communications Law Group, "New York Enacts AI Companion Safety Law Protecting Minors".
- FIRE, "California's new chatbot law has a free speech problem".
- Fortune, "New York just banned AI chatbots from pretending to be human", October 28, 2026.
We didn't read the statutes directly, and dates may change. This is information, not legal advice. This page contains affiliate links and we earn commission at no cost to you.
