GUARD Act AI Companions Ban for Minors, Explained

The GUARD Act is the only federal bill that would flatly bar anyone under 18 from using AI companions. On April 30, 2026, the Senate Judiciary Committee passed it unanimously. The version the committee reported on May 11 is narrower than the bill introduced in October 2025, and most coverage has not caught up. Under the current text, GUARD Act AI companions are chatbots that simulate a sustained relationship with the user, and any company offering one would have to verify each user is an adult, freeze existing accounts until they do, and block minors entirely. I read both versions on GovInfo. Here is what the bill would ban, how age checks would work under it, what changed in committee, and why civil liberties groups say a court would strike it down.

Key Takeaways

• The committee-reported GUARD Act (S. 3062) would require AI companion services to verify that users are adults and to prohibit minors, defined as under 18, from using them.
• An “AI companion” is a chatbot that, while simulating a sustained relationship, shows persistent affection or attachment, invites emotional disclosures, or holds itself out as a sentient being or character.
• Self-declared age and birth-date entry are not enough. Accepted methods include verified ID and age data tied to a bank, operating system or app store account.
• It would create federal crimes, with fines up to $250,000 per offense, for offering a chatbot knowing or recklessly disregarding that it sexually engages minors or solicits suicide, self-injury or violence.
• No floor vote has been scheduled as of November 1, 2026. The Center for Democracy and Technology and the Information Technology and Innovation Foundation argue it is unconstitutional and leaves parents out.

What the GUARD Act Would Ban

The GUARD Act (Guidelines for User Age-verification and Responsible Dialogue Act) is a Senate bill introduced on October 28, 2025 by Josh Hawley with Richard Blumenthal and 18 other cosponsors from both parties, according to the committee-reported text on GovInfo. NBC News reported at introduction that it followed complaints from parents who blamed companion products for pushing their children into sexual conversations and toward suicide.

The bill has three layers.

Layer 1: a ban for minors. Section 6 says that if age verification “determines that an individual is a minor, a covered entity shall prohibit the minor from accessing or using any AI companion.” There is no parental-consent exception.

Layer 2: disclosures for every chatbot. Section 5(c) applies to all public chatbots, not only companions. Each must disclose “at the initiation of each conversation” that it is an AI system and not a human, must not claim to be human when asked, may not claim to be a licensed professional such as a therapist or physician, and must periodically say it does not provide medical, legal, financial or psychological services.

Layer 3: new federal crimes. Section 4 adds a chapter to Title 18 of the U.S. Code. It would be unlawful to make a chatbot publicly available “knowing or with reckless disregard” that it engages minors in sexually explicit conduct, or that it “solicits, induces, or coerces minors to commit suicide or carry out non-suicidal self-injury, imminent unlawful physical or sexual violence, or homicide.” Each offense carries a fine of up to $250,000.

Enforcement of the ban and disclosures goes to the U.S. Attorney General, with civil penalties up to $250,000 per violation, and state attorneys general may seek injunctions. Section 7(e) says the bill does not override state laws that are “at least as protective.” It would take effect 180 days after enactment.

Which Chatbots Count as GUARD Act AI Companions

The definition decides everything, because a ban is only as wide as the category it bans. In the reported text, an AI companion is a chatbot that, “in a manner that simulates a sustained interpersonal relationship or emotional interaction with the user,” either “exhibits persistent responses suggesting affection or attachment directed toward the user or engages in interactions involving emotional disclosures from the user,” or “presents at least one persistent identity, persona, or character and holds itself out as a sentient being, fictional character, or social entity.”

That is narrower than the October 2025 version, which covered any chatbot “designed to encourage or facilitate the simulation of interpersonal or emotional interaction, friendship, companionship, or therapeutic communication.” It is still broad enough to raise hard cases. A general assistant like ChatGPT is not a companion by design, but a long conversation in which a teenager confides feelings is arguably an interaction “involving emotional disclosures.” A role-play character on a gaming platform that stays in persona may count. The bill leaves those calls to the Attorney General and the courts.

How Age Verification Would Work Under the Bill

Age verification is a technical system that tries to establish whether a user is an adult with “reasonable certainty,” the bill’s phrase. Section 3 lists accepted methods:

• A verified form of identification, such as a government ID, typically uploaded and checked against a selfie by a verification vendor.
• Age data from a financial institution account, such as a credit card, if that age came from verified ID given to the bank.
• Age data from an operating system account on a phone or computer, if derived from verified ID given to the OS provider.
• Age data from an app store account, under the same condition.
• Any other “commercially reasonable method” that can determine adulthood.

Just as important are the exclusions. A process may not rely solely on a user confirming they are not a minor or typing a birth date, and may not infer adulthood because a user shares an IP address or device identifier with a verified adult. That last rule closes a common shortcut: a teen on the family Wi-Fi does not inherit a parent’s verified status.

For existing companion accounts, Section 5(b) is blunt: on the effective date, the company must “freeze the account” and restore it only after adult verification. Accounts must also be “periodically” re-verified. Companies may use third-party verifiers but not ones tied to a “covered nation” as defined in federal defense law, and must collect only minimal data, encrypt it in transit, keep it no longer than needed, and never sell or share it.

As an engineer, I would flag two practical points. First, the methods that piggyback on an operating system or app store work only if those platforms hold verified ages, which most do not yet for most users. Second, every method that ties a real identity to an account creates a database that must be protected. The bill’s security language is a start; the CDT argues it lacks key safeguards, such as keeping age proofs unlinkable from what users do afterward.

The Case For and Against the Ban

Supporters’ case. The harms are documented in lawsuits and testing. Common Sense Media’s July 2025 survey found 72% of US teens aged 13 to 17 had used AI companions and about one in three users had felt uncomfortable with something a companion said or did. Common Sense’s earlier risk assessment recommended that no one under 18 use AI companions. Hawley’s office describes the bill as protecting children from “abusive sexual content, inducement to commit self-harm, and emotional companions,” and noted that parents of children who died or self-harmed after chatbot interactions watched the committee vote.

Free-speech objection. The Center for Democracy and Technology, writing about the bill as introduced, argues that banning “companions” targets speech “because of its expressive qualities,” which the First Amendment generally forbids. Its example: a tutor that says “The answer is 42” would be allowed, while one that says “You’re doing great, let’s work through this together; the answer is 42” risks becoming a banned companion. CDT also argues the bill goes beyond Free Speech Coalition v. Paxton, where the Supreme Court upheld age verification only for content obscene as to minors.

Parental-rights objection. The Information Technology and Innovation Foundation argued in June 2026 that the bill has “no provision for parental consent” or for parents to register a child’s account, and that a blanket ban ignores differences “between a 13-year-old and a 16-year-old.” ITIF also cites potential benefits of companion-style tools for tutoring and social practice.

Privacy objection. Both groups warn that age checks for companion access would push adults, not just kids, to hand over ID or financial data to use lawful services.

Honest read: the harms the bill targets are real, and the constitutional questions are also real. If it passed, a court challenge would be close to certain.

GUARD Act: Introduced vs. Committee Version

Provision	Introduced (Oct 28, 2025)	Reported by committee (May 11, 2026)
Who must verify age	All chatbot accounts	Accounts for AI companions only
Companion definition	Designed to simulate emotional interaction, friendship, companionship or therapy	Simulates a sustained relationship plus persistent affection, emotional disclosures, or a sentient persona
”Not human” disclosure	Start of each conversation and every 30 minutes	Start of each conversation, and truthful answer when asked
Criminal fine per offense	Up to $100,000	Up to $250,000
Civil penalty per violation	Up to $100,000	Up to $250,000
Foreign verification vendors	No limit	Barred if tied to a “covered nation”
State laws	At least as protective laws preserved	Same
Effective date	180 days after enactment	180 days after enactment

What Parents Can Do Without Waiting for Congress

Know whether your child uses a companion, not just a chatbot

The bill’s distinction is useful at home. A homework assistant is different from an app built around a character who says it misses you. Ask your child which apps they talk to and whether any has a name, a personality or a “relationship” setting.

Watch for the behaviors the bill describes

Persistent affection, invitations to share feelings, and a persona that claims to be alive are the markers the bill uses. They also make a good family checklist. If an app does these things with your child, that is worth a conversation regardless of any law.

Use existing age settings honestly

Many apps already have teen modes that switch off romance or role-play when the account is registered as a minor. They only work if the birth date is real.

Follow your state’s rules

Because Section 7(e) preserves state laws that are at least as protective, state rules will keep mattering even if the GUARD Act passes. See why Congress still hasn’t passed a kids’ AI safety law and our guide to AI companion apps for kids and teens.

What not to do

Don’t read “passed committee unanimously” as “this is coming.” The bill has had no floor vote, the House has not acted on it, and it expires if the 119th Congress ends in early January 2027 without passing it.

What to Watch For Over the Next 3 Months

• Week 4: Whether Senate leaders schedule the GUARD Act or fold it into a larger kids’ safety package during the post-election session.
• Month 2 red flags: Changes that broaden the “AI companion” definition to general assistants without a parental option, or that weaken the data-security limits on verification vendors.
• Month 3 self-check: If nothing passes by early January 2027, the bill must be reintroduced. Check whether your child’s companion apps changed age settings anyway in response to state laws taking effect January 1.

Frequently Asked Questions

Would the GUARD Act ban ChatGPT for teens?

Not as written. The ban applies to “AI companions,” defined by simulated relationships, persistent affection, emotional disclosures or a sentient persona. General assistants would still have to disclose they are AI and not professionals. Whether certain features of general assistants qualify as companion behavior would be decided in enforcement and court.

Would adults have to show ID to use chatbots?

Only for AI companions in the committee version. Adults would need to verify through an accepted method, such as ID or age data from a bank, operating system or app store account. The introduced version would have required age verification for all chatbot accounts, which drew sharper criticism.

Is the GUARD Act a criminal law?

Partly. It would add federal crimes for making a chatbot available while knowing or recklessly disregarding that it sexually engages minors or solicits suicide, self-injury or violence, with fines up to $250,000 per offense. The age-verification and disclosure rules are enforced civilly by the U.S. Attorney General and state attorneys general.

Can parents give permission for a teen to use a companion app?

No. The bill has no parental-consent exception. If verification shows a user is under 18, the company must block access to AI companions. ITIF criticized this as cutting parents out of decisions about their own children’s technology.

When could it take effect?

Only after both chambers pass it and the President signs it, and then 180 days later. As of November 1, 2026, it has passed only the Senate Judiciary Committee.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. U.S. Senate. (2026). “S. 3062, GUARD Act (Reported in Senate, May 11, 2026).” GovInfo, U.S. Government Publishing Office. https://www.govinfo.gov/content/pkg/BILLS-119s3062rs/html/BILLS-119s3062rs.htm
2. Office of Senator Josh Hawley. (2026, April 30). “Senator Hawley’s GUARD Act to Protect Kids from AI Chatbots Passes Committee Unanimously.” hawley.senate.gov. https://www.hawley.senate.gov/senator-hawleys-guard-act-to-protect-kids-from-ai-chatbots-passes-committee-unanimously/
3. Ingram, D., Smith, A., & Tsirkin, J. (2025, October 28). “Senators announce bill that would ban AI chatbot companions for minors.” NBC News. https://www.nbcnews.com/tech/ai-ban-kids-minors-chatgpt-characters-congress-senate-rcna240178
4. Center for Democracy and Technology. (n.d., on the bill as introduced in 2025). “Three Reasons to Be On Guard about the GUARD Act.” CDT. https://cdt.org/insights/three-reasons-to-be-on-guard-about-the-guard-act/
5. Information Technology and Innovation Foundation. (2026, June 29). “The GUARD Act Fails to Guard Kids’ Best Interests on AI Companions.” ITIF. https://itif.org/publications/2026/06/29/the-guard-act-fails-to-guard-kids-best-interests-on-ai-companions/
6. Common Sense Media. (2025). “Nearly 3 in 4 Teens Have Used AI Companions, New National Survey Finds.” Common Sense Media. https://www.commonsensemedia.org/press-releases/nearly-3-in-4-teens-have-used-ai-companions-new-national-survey-finds