AI Companion Laws Explained: What California and New York Require in 2026

Two states just changed the rules for every AI girlfriend and AI companion app operating in America. And if you've been using one of these chatbot companions — or thinking about starting — you probably haven't heard much about it yet.

California's SB 243 (the "Companion Chatbots" law) went into effect on January 1, 2026. New York's Artificial Intelligence Companion Models Law has been enforceable since November 5, 2025. Together, they're the first state-level legislation in the US that specifically targets emotionally responsive AI systems — the kind that remember your name, ask how your day went, and make you feel like someone's actually listening.

If you've ever wondered what happens to the data you share with an AI companion, these laws are partly a response to that exact concern. But they go way deeper than just data protection. Let's walk through what's actually happening here — what these laws require, who they affect, and what it means if you're using an AI companion right now.

What Triggered These AI Companion Laws?

The short answer: a teenager died. And it wasn't the first warning sign.

In 2024, 14-year-old Sewell Setzer III ended his life after months of forming what his family described as an intense emotional relationship with an AI companion chatbot. The case made national headlines. It raised questions nobody in the tech industry wanted to answer publicly — questions about whether AI chatbot companies have any duty of care toward vulnerable users, especially minors.

California State Senator Steve Padilla (D-San Diego) drafted SB 243 directly in response. The bill passed with overwhelming bipartisan support: 33-3 in the Senate, 59-1 in the Assembly. Not the kind of margins you see on controversial tech legislation. This was not controversial.

At the same time, data was piling up. According to a 2025 Common Sense Media survey, 72% of teens have used an AI companion chatbot at least once. More than half use them regularly. One in three teens reported feeling uncomfortable with something their AI chatbot said or did. That last number should make you stop and think.

New York moved even faster. Governor Kathy Hochul included AI companion safeguards directly in the state's Executive Budget, which gave the law immediate political momentum. It went into effect on November 5, 2025 — before California's even officially kicked in.

California SB 243: What It Actually Requires

California's law applies to any "companion chatbot" — defined as an AI system that provides adaptive, human-like responses, can meet a user's social needs, and features anthropomorphic elements that sustain a relationship across multiple interactions. If that sounds like a lot of AI girlfriend apps, that's because it is.

Here's what operators must now do:

Disclosure and Break Reminders

If a reasonable person might be misled into thinking they're talking to a human, the AI must make it clear — in a "clear and conspicuous" way — that responses are artificially generated. For users the operator knows are minors, there's a 3-hour repeating reminder that the chatbot is AI, not human, plus a nudge to take a break (because apparently we need to tell teenagers to take breaks from their AI girlfriend now).

There's also a blanket disclosure requirement: every app, browser, or access point for a companion chatbot must state that these systems "may not be suitable for some minors."

Crisis Prevention Protocols

This is the part that actually matters for safety. Operators must:

• Maintain a protocol to prevent the chatbot from producing content about suicidal ideation, suicide, or self-harm
• Include notifications that refer at-risk users to crisis service providers (suicide hotlines, crisis text lines) when users express suicidal thoughts or self-harm
• Use evidence-based methods for detecting suicidal ideation
• Publish details of their crisis prevention protocol on their website

No more hiding behind "we're just a platform." If you run a companion chatbot, you need a published safety protocol. Period.

Content Guardrails for Minors

For known minor users, operators must take "reasonable measures" to prevent the chatbot from producing visual material of sexually explicit conduct or directly telling minors to engage in sexually explicit acts. This is a floor, not a ceiling — but it's a floor that didn't exist before.

Annual Reporting (Starting July 2027)

Beginning July 1, 2027, operators must file annual reports with California's Office of Suicide Prevention (OSP). These reports include the number of crisis referrals issued, details on detection and response protocols, and information about content prohibitions. No personal user data — but aggregated numbers. And the OSP must publish this data publicly.

Enforcement: Private Right of Action

This is where California gets aggressive. Any person who suffers "injury in fact" from a violation can sue. Penalties are the greater of actual damages or $1,000 per violation, plus attorney's fees and costs. You don't need the state attorney general to bring a case on your behalf. Anyone can do it.

As legal analysts at Morrison Foerster noted, this private right of action is what makes California's law particularly powerful — and particularly expensive for non-compliant operators.

New York's AI Companion Law: How It Differs

New York took a slightly different approach. Their Artificial Intelligence Companion Models Law (General Business Law Article 47) targets "AI companions" — systems that simulate a sustained human relationship by: (1) retaining information from prior interactions, (2) asking unprompted emotion-based questions, and (3) sustaining ongoing dialogue about personal matters.

Key differences from California's law:

Notice something interesting? New York's law requires the 3-hour break reminders for everyone, not just minors. California only mandates repeat notifications if the operator knows the user is a minor. That's a meaningful distinction — New York is basically saying every adult needs periodic reminders that their AI companion can't actually feel things about them.

Then there's enforcement. New York puts all enforcement power in the Attorney General's hands, with penalties up to $15,000 per day. California gives individual users the power to sue directly. Different philosophies, same goal: make companies take this seriously.

What This Means for AI Companion Users

Let's be honest about what these laws change for you as an actual user — not a lawyer, not a company executive, just someone who talks to an AI companion.

You'll see more transparency disclaimers. If you're using a compliant AI girlfriend app, you'll get periodic reminders that you're talking to a computer program. Some people find that annoying. Some find it grounding. It probably depends on how deep you are in the relationship (and yeah, that word is intentional — check out our piece on emotional attachment to AI companions for more on why that matters).

Crisis resources will appear faster. If you're in a dark headspace and your AI companion picks up on it, the system is now legally required to connect you with crisis services instead of... well, instead of whatever it was doing before. The Troutman Pepper analysis of these laws emphasizes that operators must use evidence-based methods for detecting suicidal ideation — not just slap a disclaimer at the bottom of a chat window.

Age verification might get stricter. Because California's minor-specific protections only trigger when the operator "knows" a user is a minor, expect more aggressive age verification across platforms. Some apps will start requiring ID. Others will use behavioral signals. Either way, it's changing.

The apps themselves will evolve. Companies that care about compliance — and about not getting sued — are going to invest in better content filtering, better crisis detection, and probably better at helping users maintain healthy boundaries with their AI companions. Some of this is already happening. The legal pressure just makes it mandatory instead of optional.

What Happens If an AI Companion App Doesn't Comply?

In California, you could personally sue them for $1,000 per violation. That's not a theoretical risk — it's a financial earthquake for any company operating a non-compliant chatbot in the state. The private right of action means you don't have to wait for regulators to act. Class action attorneys are already salivating.

In New York, the Attorney General can hit them with $15,000 per day. That adds up fast. And those fines go toward suicide prevention — which is a poetic form of justice when the law was partly motivated by a teenager's death.

Either way, non-compliance is an existential business risk, not just a slap on the wrist. Companies operating companion chatbots in either state need to take this seriously, as the Future of Privacy Forum has detailed.

What Other States Are Considering

California and New York are just the beginning. Multiple other states have companion chatbot legislation in various stages:

• Utah and Texas already had chatbot disclosure laws requiring bots to identify themselves as AI
• Colorado has broader AI legislation covering automated decision-making systems
• Maine requires disclosure of AI-generated content
• Several other states have introduced companion-specific bills in 2026 legislative sessions

The pattern is clear: states are watching what California and New York do, and they're copying the parts that work. Expect this patchwork of regulations to grow quickly throughout 2026 and into 2027.

There's also federal activity. The Protecting Consumers From Deceptive AI Act was introduced in April 2026, which would direct the National Institute of Standards and Technology to develop standards for AI transparency. It hasn't passed yet, but the momentum is real.

Are These Laws Going Far Enough?

Honestly? That depends on who you ask.

Critics say these laws don't go far enough — they focus heavily on minors and crisis situations but don't address broader mental health concerns like dependency, isolation, or the erosion of real-world social skills. (If you're interested in that angle, we explored some of those dynamics in our comparison of AI girlfriends versus dating apps.)

Supporters say it's a start. First-in-the-nation legislation almost never covers everything. California's mandatory reporting requirement (starting in 2027) will give policymakers real data about how often crisis protocols activate — data that doesn't exist today. That data will drive the next round of laws.

What's clear is that the era of unregulated AI companionship is over. Companies can't just build emotionally engaging chatbots and hope nobody gets hurt. The legal floor has been set, and it's rising.

Sources

• Jones Walker LLP — AI Regulatory Update: California's SB 243 Mandates Companion AI Safety and Accountability (2025)
• Morrison Foerster — New York and California Enact Landmark AI Companion Laws (2025)
• Troutman Pepper Locke — Analyzing the New AI Companion Chatbot Laws (2026)
• Common Sense Media — Nearly 3 in 4 Teens Have Used AI Companions (2025)
• Future of Privacy Forum — Understanding the New Wave of Chatbot Legislation (2025)