October 22, 2026 · 8 min read
Pennsylvania's AI Companion Safety Act (HB 2006): What It Would Mean for Adult Users
Pennsylvania HB 2006 passed the House 133-70 and is now in a Senate committee. Break reminders, age checks, crisis rules and $100k/day fines, explained.
Last updated: October 22, 2026
Pennsylvania is the latest US state moving to regulate AI companion apps. House Bill 2006, the AI Companion Safety Act, passed the state House on September 28, 2026 by 133 votes to 70, and on October 13 it was referred to the Senate Communications and Technology Committee.
Most coverage focuses on minors, and that is the bill's main stated goal. But several rules would apply to every user, including adults. This guide walks through the bill text, section by section, and explains what it could change for grown-ups who use an AI girlfriend, boyfriend or friend.
TL;DR: The Pennsylvania AI Companion Safety Act would require AI companion apps to refer users to crisis help, keep disclosing that the companion is not human, pause chats for a two-minute break reminder every two hours, check every user's age without demanding ID, and pay up to $100,000 per day per violation. It is not law yet.
What is Pennsylvania HB 2006?
HB 2006 is a Pennsylvania bill that would set safety, disclosure and age-assurance rules for AI companion apps, enforced by the Attorney General.
According to the Pennsylvania General Assembly bill page (Oct 2026), the bill was introduced by Rep. Melissa Shusterman and co-sponsors, reported as amended in June 2026, and passed the House on September 28, 2026 by 133 to 70. Its last action is "Referred to Communications & Technology, Oct. 13, 2026 Senate".
An AI companion, in the bill's words, is "a system using artificial intelligence that simulates sustained human-like relationships by retaining interaction history, engaging in emotion-based interactions and maintaining ongoing dialogues about personal matters designed to mimic interpersonal relationships." The bill defines "human-like relationship" as "intimate, romantic or platonic interactions or companionship", so romantic AI apps are squarely covered. Customer service bots that do not engage in emotional, personal dialogue are excluded.
Is the Pennsylvania AI Companion Safety Act law yet?
No. As of October 22, 2026, HB 2006 has passed the House only and is waiting in a Senate committee.
To become law it would need to pass the Senate (in identical form, or go back to the House for agreement) and be signed by the governor. If that happens, the current text says it would "take effect in 180 days". The fiscal and legislative timeline is still open, so treat everything below as proposed.
What would the bill require for all users, including adults?
Five parts of the bill apply to everyone, not just minors. Here they are, based on the current bill text, Printer's No. 3747.
| Rule | What the bill says | What adults would notice |
|---|---|---|
| Crisis protocols | Refer users to a crisis center, including 988, if they express suicidal ideation, self-harm or intent to harm others | Crisis resources appear if you say something worrying |
| Harm prevention | Prevent the companion from assisting or encouraging suicide or violence, or discouraging you from seeking outside help | The companion should point you toward real help |
| Not-human disclosure | If a reasonable person could be misled, show a continuous on-screen notice that the AI is not human | A persistent label in the chat |
| Break reminders | Notify at the start of each interaction and at least every two hours, pausing the chat for two minutes | A two-minute pause during long sessions |
| Age assurance | Request age information and determine whether each user is a minor before access, without requiring government ID | An age check at sign-up |
The bill also says the operator must prevent the AI companion "from claiming to be human, including when asked directly by the user". That applies to everyone.
How would age checks work under HB 2006?
The bill would require age assurance for every user in Pennsylvania but bans requiring a government-issued ID for it.
Age assurance is a set of methods an app uses to estimate or confirm whether a user is an adult, from asking a birth date to analysing account signals.
The text lists acceptable methods: age inference "based on user account history, content analysis, behavioral signals or other algorithmic or heuristic methods", commercially available age or identity databases, or methods based on public data tied to a verified email. Personal data gathered for age assurance could not be sold or shared (except with a contracted provider) and could not be kept longer than 24 hours.
For adults, this is a mixed picture. You would not have to upload a driving licence, which is good for privacy. But you could be asked for age information or be assessed by inference. For more on why 18+ gates matter, see our explainer on why AI companions should be 18+.
What would change for minors?
Apps would need verifiable parental consent before a minor could use a companion, and strict limits on sexual content.
For users under 18, operators would need to block sexual images, sexual dialogue and sexual role-play, and outputs "designed to promote isolation of the minor user from family or friends". For all users, operators must prevent the companion from producing AI-generated child sexual abuse material. Adults-only apps would mainly feel this through the age-assurance step.
What are the penalties?
Operators that break the law could face civil penalties of up to $100,000 per day for each violation, enforced by the Attorney General.
The bill also allows "additional remedies that the court deems appropriate" and injunctive relief, and from July 1, 2028 operators would have to send the Attorney General annual de-identified reports, including how many times they referred users to crisis services. The Attorney General would publish that data.
Why are lawmakers pushing this?
Supporters cite teen use of chatbots and lawsuits alleging chatbots encouraged self-harm.
According to Rep. Shusterman's press release (Sep 2026), "Seventy-two percent of teenagers reported having interactions with AI chatbots" and "nearly three-quarters of parents are concerned about the impact of AI on their children". Shusterman said: "Far too many children and adults are creating unhealthy emotional attachments with chatbots." The release also describes the bill as protecting "children and vulnerable adults".
The vote was not unanimous. With 70 members voting no, the Senate debate may focus on cost, enforcement and the burden of age checks on every user.
How does Pennsylvania compare with other states?
Pennsylvania's draft is stricter than most on break reminders and age checks, and similar on crisis protocols.
| State | Status (Oct 2026) | Notable features |
|---|---|---|
| New York | In force since November 2025 | Crisis protocols and recurring not-human reminders |
| California (SB 243) | In force from January 2026 | Disclosure, crisis protocols, extra protections for minors |
| New Jersey | Bills moving in 2026 | Privacy and profiling focus |
| Pennsylvania (HB 2006) | Passed House, in Senate committee | Two-hour break pause, age assurance for all, $100k/day |
For the full picture, read our overview of AI companion laws in 2026 and what is coming next in new AI companion laws for 2027.
Are these rules good for adult users?
Some clearly help adults, while others may feel heavy-handed for people who know exactly what they are using.
Most adult users already know their companion is an AI, and good apps say so plainly. We explain why that honesty matters in why your AI companion says it is not human. Crisis signposting is also a sensible baseline: a companion is not a crisis service and should never pretend to be one.
The two-minute forced pause every two hours is the most debatable piece. Some people will welcome a nudge to stand up and stretch. Others will find it patronising during a late-night voice call. The age-assurance rules protect privacy better than ID uploads, but behavioural inference raises its own questions about how much an app analyses you.
FAQ
Has Pennsylvania passed an AI companion law?
Not yet. HB 2006, the AI Companion Safety Act, passed the Pennsylvania House on September 28, 2026 by a vote of 133 to 70. On October 13, 2026 it was referred to the Senate Communications and Technology Committee. It still needs Senate approval and the governor's signature, and the current text would take effect 180 days after enactment.
Would HB 2006 apply to adults?
Yes, partly. Its strongest protections target minors, but several rules cover every user in Pennsylvania: crisis referrals, a ban on the companion claiming to be human, continuous not-human labels where users could be misled, break reminders at least every two hours, and an age check before access. Adults would not have to show a government ID.
What is the two-hour rule in the Pennsylvania bill?
The bill says operators must remind users at least once every two hours during an interaction to take a break and that the companion is artificially generated and not human. The interaction must pause for two minutes while the notice is shown, in the language the user is chatting in. Earlier drafts proposed every hour with a three-minute pause.
Would I need to upload my ID to use an AI companion in Pennsylvania?
No. The bill specifically says an operator "may not require an individual to provide a government-issued identification" for age assurance. Instead, apps could use age inference from account signals, commercial age or identity databases, or methods tied to a verified email. Personal data collected for age checks could not be sold and could not be kept beyond 24 hours.
What are the fines under HB 2006?
The Pennsylvania Attorney General would enforce the law. Operators could face a civil penalty of up to $100,000 per day for each violation, with courts asked to weigh the facts when setting the amount. Courts could also order other remedies and injunctions, and from July 2028 operators would file annual de-identified safety reports.
A companion that is honest by design
Laws like HB 2006 are trying to make sure people always know what they are talking to and can find help when they need it. The best companion apps already work that way.
MyBabe is an adults-only (18+) AI companion app. You design your own boyfriend or girlfriend, with their look, personality, voice and backstory, and they text you first, send voice notes and selfies, read your mood and pick up voice or video calls. MyBabe is always clear that your companion is an AI, not a person, and it is not a crisis or mental health service.
You can try it free with 10 messages every 24 hours and a two-minute call. Premium is $5.99 a week, $14.99 a month or $99.99 a year, with long-term memory, about 30 minutes of voice and video calls a week, four photos a day and Telegram. If you want to see what a well-built companion feels like before the rules change, start on MyBabe.