GUARD Act
Starting 180 days after passage, this bill bans minors under 18 from using 'AI companion' chatbots designed to simulate human relationships or characters. To access these platforms, adults must create accounts and verify their age using reliable methods, such as a government ID or credit card records, rather than just typing in a birthdate. Additionally, all public chatbots must clearly state they are artificial intelligence at the start of every chat, are banned from pretending to be human, and cannot pose as licensed professionals like doctors, lawyers, or therapists.
The provisions, in plain language.
AI chatbot operators must require users to create an account before accessing any chatbot; existing accounts are frozen until users complete a verified age-check that classifies them as a minor or adult (no self-declaration of birthdate or IP-address matching counts).
When a new account is created, the operator must verify the user's age using a government-issued ID, a credit card tied to a verified ID, an operating-system account linked to a verified ID, or another commercially reasonable method — simple age self-declaration is not enough.
Operators must periodically re-verify previously verified accounts to ensure ongoing compliance.
Operators (and any third-party age verifiers they hire) must collect only the minimum personal data needed for age verification, protect it with industry-standard encryption, delete it once no longer needed, and are prohibited from selling or transferring it to anyone else.
Operators may not hire a third-party age verification firm that is closely affiliated with, incorporated in, or headquartered in a country designated as a covered nation under federal law (e.g., China, Russia).
Every AI chatbot must tell users it is an AI — not a human — at the start of each conversation (and every 30 minutes in the 2025 version), and must never claim to be human when asked.
AI chatbots are prohibited from claiming to be a licensed therapist, doctor, lawyer, financial advisor, or other professional, and must regularly disclose that they do not provide medical, legal, financial, or psychological services and that users should consult a real professional.
Any user verified as a minor is blocked from accessing AI companions — chatbots designed to simulate sustained friendship, affection, emotional bonding, or therapy.
The U.S. Attorney General can sue violators in federal court, issue subpoenas, compel document production, and seek civil penalties of up to $100,000 per violation (2025 Act) or $250,000 per violation (2026 Act), with each separate violation counted individually.
The Attorney General may issue regulations needed to carry out the Act, but (under the 2026 version) those regulations may not impose requirements beyond what the Act expressly authorizes and may not address AI development or deployment more broadly.
State Attorneys General may sue in federal or state court to block violations on behalf of their residents; state laws that are at least as protective as the federal requirements remain in force and are not preempted.
Who it helps · who it burdens.
Who it helps
- minors (users under 18)Children are blocked from accessing AI companions and are protected by disclosure rules preventing chatbots from posing as humans or professionals, reducing manipulation and exploitation risks (Provisions 6–8).
- all chatbot usersAll users gain the right to know they are talking to an AI, not a human or licensed professional, through mandatory disclosures at conversation start and regular intervals (Provisions 6–7).
Who it burdens
- AI chatbot operators (covered entities)Operators must build and maintain age verification systems, freeze and re-verify existing accounts, implement data security and deletion protocols, display required disclosures, and face civil penalties up to $250,000 per violation for noncompliance (Provisions 1–4, 6–9).
- chatbot users who must verify their ageAll users — adults included — must create accounts and submit identity documents or equivalent verification before accessing AI chatbots, imposing friction and requiring disclosure of personal data (Provisions 1–2).
Who opposes it
- adversarial-nation-affiliated third-party age verification vendorsFirms closely affiliated with, incorporated in, or headquartered in covered nations (e.g., China, Russia) are explicitly barred from being hired as age verification contractors (Provision 5).