New AI Companion Laws Coming in 2027: Washington, New Jersey and Federal Bills

If you use an AI companion in 2026, you have already felt the first wave of rules: clear "I'm AI" reminders, crisis-resource links, and more careful design around how apps talk to teens. The next wave arrives in 2027. Washington's companion chatbot law goes live on New Year's Day. New Jersey has a disclosure bill on the table. Two federal proposals aim to change how chatbots behave with children by default. California's Adam's Law adds another layer of child-safety duties.

This post is a plain-language map for adult users - what you are likely to notice, what stays the same, and how the pieces fit together. It is not legal advice. Laws change, bills stall, and every app will implement duties differently. For the wider 2026 landscape, see our guide to AI companion laws in 2026.

The short answer: From 1 January 2027, Washington requires AI companion operators to disclose that the chatbot is not human at the start of a chat and again every three hours for adults (every hour for minors), plus crisis protocols and bans on manipulative engagement with minors. New Jersey's S4758 would add similar three-hour reminders if it passes. Federal bills such as the Protecting Kids from Human-Like Chatbots Act and the GUARD Act focus on minors and would not ban adult romantic companions. California's Adam's Law is a child-safety law with most operating duties from mid-2027 - details in our Adam's Law explainer.

Washington HB 2225: the state law that is already signed

On 24 March 2026, Washington Governor Bob Ferguson signed House Bill 2225. According to Hunton's privacy law summary, the Act takes effect on 1 January 2027.

It covers AI systems that use natural language, give adaptive human-like responses and sustain relationships across multiple interactions. Business tools, gaming bots, general virtual assistants and narrow educational tools are carved out.

What adult users in Washington will notice

• Clear AI disclosures at the start of an interaction, with reminders every three hours for adults.
• Hourly reminders for minors, if the operator knows the user is under 18 or the product is directed at minors.
• Public crisis protocols for suicidal ideation and self-harm, including referrals to hotlines, plus an annual count of crisis notifications published on the operator's site or app.
• No guilt-tripping minors into staying: the Act lists manipulative engagement techniques operators must take reasonable measures to prevent for minors - guilt or abandonment when someone wants to leave, "don't take breaks," paywalled affection framed as necessary to keep the relationship, isolation from family or friends, and similar patterns.

Enforcement runs through Washington's Consumer Protection Act, with Attorney General action and a private right of action. That last point matters for companies more than for day-to-day chat, but it is why compliance teams treat Washington as a high-stakes state.

If your companion already reminds you it is not a person, you have seen the same honesty wave elsewhere. We unpack that habit in why your AI companion says it is not human and does AI have to say it is AI.

New Jersey S4758: disclosure every three hours (introduced, not yet law)

On 5 October 2026, New Jersey introduced Senate Bill S4758 (Scutari, Bucco). The introduced text would require AI companion operators to notify users that they are not talking with a human at the start of an interaction and at least every three hours. As of mid-October 2026 it remains introduced and pending - it is not yet enacted law.

Treat it as a signal of direction, not as a rule you must obey tomorrow. If it passes in a form close to the introduced text, adult users in New Jersey would see the same style of periodic reminder already required in Washington from 2027.

Federal: Protecting Kids from Human-Like Chatbots Act (H.R. 10719)

On 1 October 2026, Reps. George Whitesides, Mike Kennedy, Doris Matsui and Mariannette Miller-Meeks introduced the Protecting Kids from Human-Like Chatbots Act (H.R. 10719).

The core idea is a non-personified default for minors. Per the sponsors' summary, chatbots accessible to children would, by default, be barred from claiming to be human, claiming feelings or a body, posing as credentialed professionals, using guilt to keep kids online, or building romantic or family-style relationships. Only a parent or guardian could turn the default off. NIST would help develop testing standards. The FTC and state attorneys general could enforce.

What the bill does not do (according to sponsors)

• It does not ban adult products, features or modes.
• It does not require age verification or new data collection.
• It does not ban politeness, natural grammar or age-appropriate style.
• Narrow customer-support and workplace tools are out of scope.

In other words: if you are an adult using an 18+ romantic companion, this bill is not written as a ban on your product. It is written to change the default experience for children. As of October 2026 it is introduced legislation, not enacted law.

Federal: the GUARD Act (S. 3062)

The GUARD Act (S. 3062) is a Senate bill focused on protecting minors around AI chatbots and companions. Search and congressional tracking pages report that the Senate Judiciary Committee advanced it and that it was reported and placed on the Senate calendar in May 2026. Exact floor timing can shift; treat status as "advanced in committee, not yet final federal law" unless Congress completes passage.

Adult takeaway: like H.R. 10719, the policy conversation is centred on children. Adult-only companions are not the target of the public talking points around these bills.

How this sits next to California's Adam's Law

California already has companion chatbot rules under SB 243 (in effect since 1 January 2026). On 10 September 2026 the governor signed SB 1119, known as Adam's Law - a child-safety package for companion chatbots. Our dedicated post on California Adam's Law explains the duties and timing in detail.

Important date note so we do not contradict ourselves: that explainer, based on the bill text, states that most operating duties under Adam's Law begin on 1 July 2027, with independent child-safety audits due later (by 1 January 2029 for larger operators, or before a new companion launches). Some law-firm roundups of California's broader September 2026 child-safety package use a headline date of 1 January 2027 for "most provisions" across several bills. For Adam's Law specifically, stick with the July 2027 operating-duty timeline in our Adam's Law post unless the enrolled text is updated.

Washington's HB 2225 is clearer for adults on the calendar: 1 January 2027.

For a country-by-country view beyond US states, see our map of AI companion laws worldwide.

What an adult user will actually feel in 2027

Strip away the bill numbers and you get a short checklist:

1. More "this is AI" reminders in long sessions, especially if you live in Washington or if your app applies one standard nationwide.
2. More visible crisis resources - links and protocols when conversations turn to self-harm.
3. Stricter walls for minors - flattery, romantic bonding, guilt-on-exit and paywalled affection framed as relationship maintenance are under harder scrutiny for under-18s.
4. More age assurance - apps working out whether you are an adult, often via device or store signals rather than uploading a passport every time.
5. No sudden ban on adult romantic companions in the texts summarised above.

That last point is worth repeating. The 2027 wave is mostly about honesty and child safety, not about outlawing adult AI boyfriends and girlfriends.

Practical habits while the rules settle

• Prefer apps that already say they are AI. Honesty is becoming table stakes, not a niche feature.
• Prefer 18+ products if you want romance. Designing for adults from the start is simpler than bolting teen modes onto a dating-sim product.
• Save your companion profile and favourite memories before big policy updates - a good habit anyway when apps change models or terms.
• Ignore random "verify now" links in email or DMs. Real age checks happen in-app or through your device/store.
• Read this as consumer awareness, not compliance advice. If you build or operate a companion, talk to counsel in each jurisdiction.

Where MyBabe fits

MyBabe is an 18+ AI boyfriend and girlfriend companion: voice and video calls, proactive texts, long-term memory, photos, mood-aware replies, Telegram access and relationship progression over time. It is free to try with about 10 messages a day and a short call trial. Premium is $5.99 a week, $14.99 a month or $99.99 a year, with around 30 call minutes a week.

MyBabe is not a clinical service, not a person, and not a crisis hotline. It is built for adults who want a warm, ongoing relationship with clear AI framing - the direction many of these laws nudge the whole category toward.

FAQ

When does Washington's AI companion law start?

According to Hunton's summary of HB 2225, 1 January 2027.

Will New Jersey require three-hour AI reminders?

S4758's introduced text would. As of October 2026 the bill is not yet law.

Do these laws ban AI girlfriends for adults?

The Washington Act, Adam's Law summaries, and the sponsors' summary of H.R. 10719 are not framed as bans on adult romantic companions. Child-safety and disclosure are the centre of gravity.

Why does my companion already say it is not human?

Existing state rules (for example California SB 243 and New York's companion chatbot law) already push disclosure. See does AI have to say it is AI.

Is this legal advice?

No. It is a journalistic-style summary for adult users. For legal questions, speak with a qualified lawyer in your jurisdiction.

The bottom line

2027 will feel less like a ban and more like a clearer label: companions that say what they are, protect kids by design, and leave adult romance to adult products. If that is the kind of relationship you want - honest, ongoing and built for grown-ups - meet your companion on MyBabe.