AI Companion Legal & Jurisdictional Guide
Read the AI companion legal guide: US state age laws (Texas HB 1181, SCOTUS Paxton 2025), 18 USC 1466A, EU AI Act, UK Online Safety Act, FTC click-to-cancel.
Three years ago this page would have been a paragraph. AI companion apps were a small category, no state had a live age-verification statute on them, and the only federal rule that mattered was the one nobody would ever cross. The 2023-2026 wave changed all of that. Louisiana Act 440 hit first, Texas HB 1181 followed, and what was a single-state experiment became a multi-jurisdiction perimeter: eighteen US states with age-verification duties, the EU AI Act, the EU DSA, the UK Online Safety Act, the FTC click-to-cancel rule, plus parallel regimes in Canada and Australia. None of it bans the category for adults. All of it pushes the burden onto the operator. The platforms in our test don't all handle this the same way.
This page is the parent topic-pillar. It maps the federal-level US statutes that bind every operator, the EU + UK + Canada + Australia regulatory architecture, and the cross-cutting compliance perimeter. The state-by-state walkthrough lives in a dedicated state guide. Federal child-safety statute deep-dive lives in our § 1466A explainer. Religious-jurisprudence treatment lives in our haram FAQ. The audience here is the adult reader asking what binds the platform I'm about to subscribe to, and what binds me? The short answer is: most of the burden sits on the operator. The long answer is below.
Are AI companion apps legal in the United States?
Yes, AI companion apps are legal for adults in the United States, with no federal statute prohibiting consensual adult use. Operators must comply with 18 USC § 1466A (no obscene minor depictions, including AI-generated), state age-verification statutes in eighteen states, FTC consumer-protection rules, and Section 230 obligations. Individual users face no criminal exposure for normal adult use.
The federal baseline is permissive for adult use. The First Amendment protects sexual expression for adults under Miller v. California (1973) and Ashcroft v. Free Speech Coalition (2002). There is no federal criminal statute reaching adult-only AI companion services as such. The operator-side perimeter is built from four federal layers: criminal law (18 USC § 1466A), FTC consumer-protection authority under Act § 5 + ROSCA, the conditional immunity of Section 230, and the state-level age-verification statutes that overlay the federal floor. Adult readers encountering a compliant mainstream AI companion product on the legal commercial market do not commit a crime by using it.
Which US states require age verification for adult AI sites?
Eighteen US states as of May 2026: Louisiana, Texas, Utah, Mississippi, Virginia, Arkansas, Montana, North Carolina, Idaho, Oklahoma, Kentucky, Indiana, Kansas, Florida, Nebraska, Alabama, Georgia, and Tennessee. Civil penalties reach $10,000 per day per violation. The duty binds platforms publishing the content, not editorial comparators. Several operators have geo-blocked entire states rather than deploy ID verification.
The legislative wave traces a clear arc. Louisiana opened with Act 440 in 2022, requiring digital ID credentials linked to the LA Wallet app. Texas followed with HB 1181 in 2023. Utah, Mississippi, Virginia, Arkansas, and Montana enacted comparable statutes through 2023. The 2024-2025 wave added North Carolina, Florida HB 3, and the rest of the current eighteen. The 2026 session is on track to add more bills now that constitutional uncertainty was resolved by the Supreme Court. The per-state walkthrough (effective dates, AG enforcement design, private rights of action, geo-block patterns) sits in our state-by-state guide; the table below is the federal-level operator-side summary.
| Statute | Citation | Year | Operator implications |
|---|---|---|---|
| 18 USC § 1466A | Obscene visual representations of minors | PROTECT Act 2003 | In force; reaches AI-generated content; absolute red line |
| 18 USC § 2256 | Definitions including AI-generated material | PROTECT Act 2003 | In force; statutory definitions used in § 1466A prosecutions |
| 18 USC § 2257 | Performer record-keeping | 1988 / amended 2006 | In force; applicability to fully-synthetic AI output is open question under regulatory rulemaking |
| 47 USC § 230 | Communications Decency Act safe harbor | 1996 | In force; narrowed by FOSTA-SESTA (2018); does not cover platform-authored AI output |
| FTC Act § 5 | Unfair and deceptive practices | 1914 / amended | In force; up to $51,744 per violation; basis for Operation AI Comply (September 2024) |
| FTC Negative Option Rule | Click-to-cancel (amended) | October 2024 | In force in stages through 2025; cancellation parity required; per-violation civil exposure |
| ROSCA | Restore Online Shoppers' Confidence Act | 2010 | In force; informed consent for negative-option features; AG state-level parallels |
| FOSTA-SESTA | Stop Enabling Sex Traffickers Act | 2018 | In force; narrows § 230 immunity for sex-trafficking-related content |
What is Texas HB 1181?
Texas HB 1181, signed June 2023, requires commercial sites where more than one-third of content is sexual material harmful to minors to verify visitor age via government-issued ID, transactional data, or a commercially reasonable method. The statute was upheld by the US Supreme Court in Free Speech Coalition v. Paxton on June 27, 2025. Civil penalties run to $10,000 per day.
The statute's operative threshold is the one-third rule. The duty attaches to commercial entities publishing material "more than one-third of which is sexual material harmful to minors." Civil penalties reach $10,000 per day of violation, plus up to $250,000 if a minor accesses material as a result of the violation. The Texas Attorney General is the sole enforcer; there is no private right of action. The Supreme Court's 6-3 decision in Free Speech Coalition v. Paxton (2025) applied intermediate scrutiny and rejected the strict-scrutiny framework of Ashcroft v. ACLU (2004), functionally green-lighting the parallel state statutes.
The practical effect for Texas residents is operator-by-operator. Some AI companion platforms integrated Yoti or Persona and remained available behind an ID-upload step. Others geo-blocked Texas at the IP level and call it done. When Stripchat review added Texas to its geo-block list in early 2024, we updated fourteen review pages within 72 hours so readers in Austin or Houston wouldn't click a CTA that would dead-end on a 403. That's the practical shape of "legal but inaccessible": the product is lawful, your state didn't ban it, and the operator chose not to serve you.
What is the UK Online Safety Act 2023?
The UK Online Safety Act 2023 imposes a statutory duty on services publishing or hosting pornographic content to use highly effective age assurance to prevent under-18 access. Ofcom's Part 5 guidance took effect July 25, 2025, listing approved methods including photo-ID matching, facial age estimation, mobile-carrier checks, and credit-card verification. Maximum fines reach £18 million or 10% of global revenue.
The Online Safety Act 2023 is the most consequential adult-platform regulation outside the US. Part 5 covers services publishing pornographic content directly; Part 3 reaches user-to-user services. Ofcom is the primary regulator, and its "highly effective age assurance" standard expressly excludes self-declaration. AI companion operators serving UK users have integrated third-party providers including Yoti, Persona, VerifyMy, and AgeChecked to meet the requirement. Non-compliance can trigger fines of £18 million or 10% of qualifying worldwide revenue.
The practical compliance posture has split the market. Mainstream AI companion platforms have built proper verification flows. A subset of smaller operators have geo-blocked the UK rather than carry the integration cost. The reader-side experience in 2026 is the same it became in Texas: load the app from a UK IP and you'll see a working ID step, a working facial-age-estimation step, or a polite withdrawal notice. The legal product is not banned for UK adults; access varies by operator commercial choice.
Does the EU DSA regulate AI companion platforms?
Yes. The EU Digital Services Act (Regulation 2022/2065) regulates AI companion platforms as hosting services. Article 28 requires proportionate measures protecting minors. Article 16 governs notice-and-action takedown. Very Large Online Platforms face additional risk-assessment duties under Articles 34-35. The European Commission's July 2025 guidance clarifies that adult platforms must deploy effective age assurance, not self-declaration.
The Digital Services Act entered into force November 2022 and applies fully to all hosting services since February 2024. Article 28 is the minor-protection cornerstone. Article 16 imposes notice-and-action procedures with reasoned decisions under Article 17. Articles 26-27 govern advertising transparency. The European Commission's July 2025 guidelines on Article 28 explicitly named adult platforms as targets and required deployment of effective age-assurance methods. Penalties for systemic infringement reach 6% of global annual turnover.
What this means for an operator with EU users is layered. The Article 28 duty is proportionate (the regulator weighs the platform's size, the audience risk, the technical state-of-the-art). The Article 16 takedown duty is procedural (every notice gets a reasoned response). The VLOP threshold (45 million monthly active EU users) doesn't catch most AI companion platforms today, but the smaller-platform duties already apply.
What does the EU AI Act require for AI companions?
The EU AI Act (Regulation 2024/1689) classifies most AI companion apps as limited-risk systems. Article 50 requires clear disclosure that users are interacting with AI, not a human. AI-generated synthetic content must be machine-readable and labeled as such. Manipulative techniques exploiting vulnerabilities are prohibited under Article 5. Penalties reach €35 million or 7% of global turnover for severe infringements. Article 50 enters application August 2026.
The AI Act is the world's first horizontal AI statute. Article 5 prohibits manipulative AI systems exploiting vulnerabilities of specific groups, including by reason of age, disability, or socio-economic situation. Article 50 transparency obligations apply to AI companion operators directly: the user must be told they are interacting with an AI, and AI-generated audiovisual content (including the synthetic image, voice, and chat output of an AI companion) must be marked machine-readable. Article 50 enters application August 2026. Severe infringements expose operators to €35 million or 7% of global annual turnover under Article 99.
The reader-facing practical change is the disclosure banner. Compliant operators serving EU users will surface a standardised "you are interacting with an AI" notice at session start, and any image, video, or voice clip generated by the system will carry machine-readable provenance metadata. This is operator perimeter, not user-side burden.
What is 18 U.S.C. § 1466A?
18 U.S.C. § 1466A is the US federal statute criminalising obscene visual representations of the sexual abuse of children. It explicitly reaches drawn, painted, sculpted, computer-generated, and AI-generated depictions. Production carries up to 30 years imprisonment; receipt or distribution carries a 5-year mandatory minimum. The statute applies regardless of whether any real child was harmed in creating the image. This is the universal red line of AI policy.
§ 1466A is the absolute red line in US AI policy. It was enacted as part of the PROTECT Act of 2003 precisely to reach virtual depictions after Ashcroft v. Free Speech Coalition (2002) struck portions of the Child Pornography Prevention Act. The statute survives an obscenity-based constitutional challenge because it incorporates the Miller v. California (1973) obscenity test. Any platform found to host § 1466A material is excluded from our editorial coverage regardless of commercial relationship. Our dedicated § 1466A explainer walks the statutory text, the PROTECT Act history, and the parallel state and international statutes in detail.
Does Section 230 protect AI companion platforms from liability?
Section 230 of the Communications Decency Act provides limited immunity for third-party user-generated content but does not shield platforms from federal criminal law including 18 USC § 1466A, intellectual property claims, or FOSTA-SESTA sex-trafficking exposure. Generative AI output may not qualify as third-party content under § 230. Recent rulings including Anderson v. TikTok (3rd Cir. 2024) signal narrowing application to algorithmic recommendations.
Section 230 is narrower than commonly believed. It does not immunise platforms from federal criminal law (the § 1466A carve-out is explicit), intellectual-property claims, or sex-trafficking liability under FOSTA-SESTA (2018). The open question for AI companion operators is whether AI-generated platform output qualifies as "information provided by another information content provider", the threshold for § 230 protection. The early academic and litigation consensus is that pure platform-authored AI output likely does not. Anderson v. TikTok (3rd Cir. 2024) further narrowed § 230 application to algorithmic curation choices, and the Supreme Court declined to disturb the holding on cert.
Most reviewers in this space won't tell you that the legal exposure for platform-authored AI content is materially different from the user-generated-content immunity the early consumer internet was built on. We will. An AI companion app that generates a sext, a synthetic image, or a roleplay response is the speaker for § 230 purposes; it does not get the same shield a forum host gets for user posts.
What does the FTC enforce for AI companion platforms?
The Federal Trade Commission enforces against unfair and deceptive practices under FTC Act § 5, including misleading subscription terms, dark-pattern checkout flows, AI capability misrepresentation, undisclosed sponsored content, and the Restore Online Shoppers' Confidence Act (ROSCA). The 2024 click-to-cancel rule and the Negative Option Rule directly affect AI companion subscription billing. Enforcement actions can reach $51,744 per violation.
The FTC is the most consistently active federal regulator for AI companion operators. Enforcement priorities include AI capability claims (the FTC's Operation AI Comply struck five companies in September 2024), subscription dark patterns, undisclosed affiliate relationships under the Endorsement Guides (the basis for our affiliate-disclosure veto on every page), and ROSCA-grade cancellation friction. Civil penalties under § 5 currently top out at $51,744 per violation as of 2025.
The Endorsement Guides matter to us specifically. Every page on this site carries an affiliate disclosure at the top because the FTC's 2023 guides require it; every CTA routes through a tracked redirect with clear destination context for the same reason. Operator-side ROSCA exposure has tightened parallel to the FTC's October 2024 click-to-cancel finalisation.
What is the FTC click-to-cancel rule?
The FTC's click-to-cancel rule, finalised October 2024 as an amendment to the Negative Option Rule, requires that any cancellation method must be at least as easy as the signup method. If users subscribed online, they must be able to cancel online with the same number of steps. The rule took effect in stages through 2025. Violations expose operators to civil penalties and consumer redress under § 19 of the FTC Act.
The amended Negative Option Rule is the consumer-protection lever most directly aimed at the AI companion subscription model. Several large operators historically required phone calls, hidden settings menus, or grief-tier "retention specialists" to cancel. The new rule makes that pattern per-violation civil exposure. Operators must also provide annual reminders for any subscription auto-renewal exceeding a year. Class-action plaintiffs have begun citing the rule alongside state UDAP statutes in the 2025-2026 subscription-cancellation litigation wave.
Are AI companion platforms regulated in Canada?
Yes. Canada regulates AI companion platforms under PIPEDA (federal privacy), provincial privacy statutes including Quebec Law 25 and BC PIPA, the Criminal Code § 163.1 (child sexual abuse material including AI-generated), and Bill S-210 (Protecting Young Persons from Exposure to Pornography Act) which mandates age verification on commercial pornography sites. The Canadian Radio-television and Telecommunications Commission may also exert jurisdiction over platform conduct.
Bill S-210 passed Senate third reading in 2023 and progressed through the House of Commons through 2024-2025. Quebec's Law 25 imposes robust privacy obligations exceeding PIPEDA. The Criminal Code § 163.1 reaches virtual depictions under the broad "any visual representation" language interpreted in R. v. Sharpe (2001). The compliance posture for Canadian operators is closer to the EU model than to the patchwork US one; federal-level rules dominate, provincial privacy layers add depth, and the criminal floor on child-protection content is uncompromising.
Are AI companion platforms regulated in Australia?
Yes. Australia's eSafety Commissioner enforces the Online Safety Act 2021, the Basic Online Safety Expectations, and industry codes including the Class 1A and 1B codes addressing synthetic minor sexual abuse material. The Restricted Access System mandates age-gating for adult content, and the December 2024 Social Media Minimum Age Bill restricts under-16 access to platforms broadly. Civil penalties reach AUD $782,500 per contravention.
Australia is a leading-indicator regulator. The eSafety Commissioner operates as a dedicated regulator with notice-and-action authority. The 2023-2024 industry codes addressing generative AI synthetic minor sexual abuse material were a global first. The Social Media Minimum Age Bill creating an under-16 platform restriction passed in December 2024 with effect in 2025-2026, a regime AI companion operators with Australian users must navigate.
| Region | Key statute | Year | Operator implications |
|---|---|---|---|
| United States | 18 USC § 1466A; FTC Act § 5; Section 230; 18 state AV statutes | 2003 / 1996 / 2022-2026 | Federal child-protection red line; consumer-protection enforcement; conditional UGC immunity; state-level age verification |
| European Union | GDPR; DSA; AI Act | 2018 / 2022 / 2024 | Privacy, content moderation, AI transparency layered; up to €35M or 7% global turnover penalties |
| United Kingdom | Online Safety Act 2023; UK GDPR; CDPA 1988 | 2023 / 2018 / 1988 | Highly effective age assurance for adult platforms; £18M or 10% revenue penalties; Ofcom regulator |
| Canada | PIPEDA; Bill S-210; Criminal Code § 163.1; Quebec Law 25 | 2000 / 2023 / 1985 / 2021 | Federal-provincial privacy stack; commercial-pornography age-verification mandate; criminal floor extends to virtual depictions |
| Australia | Online Safety Act 2021; Basic Online Safety Expectations; Industry Codes | 2021 / 2022 / 2023 | eSafety Commissioner enforcement; AUD $782,500 per contravention; industry codes on synthetic minor content |
What if I'm in a country with strict adult-content laws?
If your jurisdiction prohibits or restricts adult-content access, including Saudi Arabia, the UAE, India, Indonesia, Malaysia, Pakistan, China and others, you must not access AI companion platforms or this editorial site. Penalties range from administrative fines to imprisonment. Jurisdictional compliance is the visitor's responsibility under our terms of use. We do not maintain a country block-list at the editorial layer.
Adult-content access is criminalised or heavily restricted in dozens of jurisdictions. India regulates under the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules 2021 with intermediary takedown duties. China blocks adult platforms entirely under the Cybersecurity Law and the Personal Information Protection Law. Gulf states impose criminal penalties on possession or transmission. Indonesia's Pornography Law (UU 44/2008) reaches digital adult content. Local counsel is essential; penalties can include imprisonment. The religious-jurisprudence dimension for Muslim-majority readers lives in our haram-question FAQ.
Can I be charged for AI-generated content depicting minors?
Yes, in every major jurisdiction. US 18 USC § 1466A and § 2256 (PROTECT Act of 2003) reach AI-generated material. UK Protection of Children Act 1978 covers pseudo-photographs. EU Child Sexual Abuse Directive 2011/93/EU and the May 2024 amendment criminalise AI-generated child sexual abuse material. Canada Criminal Code § 163.1 covers virtual depictions. Australia's Online Safety Act addresses synthetic material. There is no jurisdiction where this is lawful.
This is the universal red line of AI policy. The UK Protection of Children Act 1978 was amended by the Criminal Justice and Public Order Act 1994 to include "pseudo-photographs" and applied to AI by the Online Safety Act 2023. The EU Child Sexual Abuse Directive 2011/93/EU was amended in May 2024 to expressly cover AI-generated material. Operators that fail to filter AI-generated material involving minors face existential criminal exposure. Users who request such content commit a federal crime. The substantive deep-dive on this question lives in our § 1466A explainer; the universal framing is in our minor-protection pillar.
What about marriage to an AI?
No US state, EU member state, or country recognises marriage to an AI. Marriage statutes universally require two human parties with legal capacity. Some operators have promoted symbolic ceremonies inside their apps, but these are not civil marriage. The 2018 Akihiko Kondo Japanese hologram ceremony and 2024 Replika community ceremonies illustrate cultural interest, not legal recognition. Future change would require legislative action.
The cultural fascination with AI marriage is real but legally inert. Every marriage statute we have reviewed (US Uniform Marriage and Divorce Act jurisdictions, UK Marriage Act 1949, French Code civil Article 144, Japan's Civil Code Article 731) defines marriage as between human persons of legal age and capacity. The 2018 Akihiko Kondo ceremony with the Vocaloid hologram Hatsune Miku and the 2024 Replika community ceremonies are explicitly non-legal. We will not speculate on whether jurisdictions might extend marriage law in the future; that question is legislative. Our forward-looking FAQ on the question walks the legislative outlook + cultural context in more depth.
How do I find a lawyer for AI-companion-related questions?
For US matters, contact your state bar association lawyer-referral service or use Martindale-Hubbell, Avvo, or the American Bar Association directory. EU residents can use the European e-Justice Portal. UK residents can use the Law Society's Find a Solicitor tool. Specialists handling AI, technology, consumer-protection, and First Amendment matters are most relevant. Our editorial team cannot recommend specific attorneys.
The US directories above are well-established starting points. The American Bar Association lists state-by-state lawyer-referral services. The European e-Justice Portal routes to national bar associations. The UK Law Society offers a public solicitor finder. For our specific subject matter, look for combinations of technology / AI / data privacy practice, First Amendment or media law (US), consumer-protection or class-action defense (operator-side), and child-safety expertise. Pro bono technology-policy organisations including the Electronic Frontier Foundation and the ACLU maintain directories of cooperating attorneys for First Amendment matters.
When this page does not apply
This page covers the federal-level US statutes and the EU + UK + Canada + Australia regulatory architecture for adult AI companion platforms as of May 2026. It does not cover: minor users (always illegal regardless of jurisdiction, see our minor-protection pillar), AI-generated content involving minors (a federal felony under § 1466A everywhere, see our § 1466A explainer), commercial operators evaluating compliance design (consult a qualified regulatory attorney), per-state US walkthrough (see our state-by-state guide), or religious-jurisprudence treatment (see our haram FAQ).
The regulatory landscape changes faster than any annual editorial cycle can track. We re-verify quarterly; readers depending on the page for compliance decisions should treat it as a starting point and consult a qualified attorney for the specific question they face.
Sources
- [Source: 18 USC § 1466A, Obscene visual representations of the sexual abuse of children · verified 2026-05-26]
- [Source: 18 USC § 2256, Definitions for child-exploitation chapter · verified 2026-05-26]
- [Source: 18 USC § 2257, Record-keeping requirements · verified 2026-05-26]
- [Source: 47 USC § 230, Communications Decency Act safe harbor · verified 2026-05-26]
- [Source: Free Speech Coalition v. Paxton (US Supreme Court, 2025) · verified 2026-05-26]
- [Source: Miller v. California (US Supreme Court, 1973) · verified 2026-05-26]
- [Source: Ashcroft v. American Civil Liberties Union (US Supreme Court, 2004) · verified 2026-05-26]
- [Source: Ashcroft v. Free Speech Coalition (US Supreme Court, 2002) · verified 2026-05-26]
- [Source: UK Online Safety Act 2023 · verified 2026-05-26]
- [Source: UK Protection of Children Act 1978 · verified 2026-05-26]
- [Source: EU Digital Services Act (Regulation 2022/2065) · verified 2026-05-26]
- [Source: EU AI Act (Regulation 2024/1689) · verified 2026-05-26]
- [Source: EU GDPR (Regulation 2016/679) · verified 2026-05-26]
- [Source: EU Child Sexual Abuse Directive (parent EU law tree) · verified 2026-05-26]
- [Source: FTC Negative Option Rule (final 2024), click-to-cancel · verified 2026-05-26]
- [Source: FTC Operation AI Comply (September 2024 announcement) · verified 2026-05-26]
- [Source: Canadian Bill S-210, Protecting Young Persons from Exposure to Pornography Act · verified 2026-05-26]
- [Source: eSafety Commissioner (Australia, Online Safety Act 2021) · verified 2026-05-26]
- [Source: Cyberspace Administration of China (regulator of adult platform access) · verified 2026-05-26]
- [Source: European e-Justice Portal (lawyer referral, EU residents) · verified 2026-05-26]
- [Source: Law Society of England and Wales (lawyer referral, UK residents) · verified 2026-05-26]
Cite this page
Joly, A. (2026). AI companion legal & jurisdictional guide. bestgirlfriend.ai. Retrieved from https://bestgirlfriend.ai/safety/ai-companion-legal
Frequently asked questions
Are AI companion apps legal in the United States?
Yes, AI companion apps are legal for adults in the United States, with no federal statute prohibiting consensual adult use. Operators must comply with 18 USC § 1466A, state age-verification statutes in eighteen states, FTC consumer-protection rules, and Section 230 obligations. Individual users face no criminal exposure for normal adult use.
Which US states require age verification for adult AI sites?
Eighteen states as of May 2026: Louisiana, Texas, Utah, Mississippi, Virginia, Arkansas, Montana, North Carolina, Idaho, Oklahoma, Kentucky, Indiana, Kansas, Florida, Nebraska, Alabama, Georgia, and Tennessee. Civil penalties reach $10,000 per day per violation. Several operators have geo-blocked entire states rather than deploy ID verification. The state-by-state walkthrough sits at our state guide.
What is Texas HB 1181?
Texas HB 1181 (September 2023) requires commercial sites where more than one-third of content is sexual material harmful to minors to verify visitor age via government-issued ID or a commercially reasonable method. Upheld by the US Supreme Court 6-3 in Free Speech Coalition v. Paxton on June 27, 2025. Civil penalties up to $10,000 per day plus up to $250,000 per minor-access incident.
What is the UK Online Safety Act 2023?
The UK Online Safety Act 2023 imposes a statutory duty on services publishing or hosting pornographic content to use highly effective age assurance. Ofcom's Part 5 guidance took effect July 25, 2025. Maximum fines reach £18 million or 10% of global revenue. Self-declaration is expressly excluded.
Does the EU DSA regulate AI companion platforms?
Yes. The EU Digital Services Act regulates them as hosting services. Article 28 requires proportionate measures protecting minors. Article 16 governs notice-and-action takedown. The European Commission's July 2025 guidance requires effective age assurance, not self-declaration. Penalties up to 6% of global turnover.
What does the EU AI Act require for AI companions?
The EU AI Act (2024/1689) classifies most AI companions as limited-risk. Article 50 requires disclosure that the user is interacting with AI; AI-generated content must be machine-readable. Article 5 prohibits manipulative AI exploiting vulnerabilities. Article 50 enters application August 2026. Penalties up to €35 million or 7% of global turnover.
What is 18 U.S.C. § 1466A?
The US federal statute criminalising obscene visual representations of the sexual abuse of children, including AI-generated. Production carries up to 30 years; receipt or distribution a 5-year mandatory minimum. Applies regardless of whether any real child was harmed. Our dedicated explainer walks the statute in full.
Does Section 230 protect AI companion platforms?
Limitedly. § 230 immunity covers third-party content but not federal criminal law including § 1466A, IP claims, or FOSTA-SESTA. Generative AI output may not qualify as third-party content. Anderson v. TikTok (3rd Cir. 2024) further narrowed application to algorithmic curation.
What does the FTC enforce?
FTC Act § 5 (unfair / deceptive practices), the Negative Option Rule, ROSCA, the Endorsement Guides, and Operation AI Comply (September 2024). Targets include misleading subscription terms, dark-pattern checkout, AI capability misrepresentation, undisclosed sponsored content. Civil penalties up to $51,744 per violation.
What is the FTC click-to-cancel rule?
Finalised October 2024 as an amendment to the Negative Option Rule. Requires that any cancellation method be at least as easy as the signup method. If you subscribed online, you must be able to cancel online with the same number of steps. Took effect in stages through 2025.
Are AI companions regulated in Canada?
Yes. PIPEDA (federal privacy), Quebec Law 25, BC PIPA, Criminal Code § 163.1 (virtual depictions included), and Bill S-210 mandating age verification on commercial pornography sites. The CRTC may also exert jurisdiction over platform conduct.
Are AI companions regulated in Australia?
Yes. The Online Safety Act 2021, the Basic Online Safety Expectations, and industry codes addressing synthetic minor sexual abuse material. Civil penalties up to AUD $782,500 per contravention. The December 2024 Social Media Minimum Age Bill restricts under-16 access broadly.
What if I'm in a country with strict adult-content laws?
You must not access AI companion platforms or this editorial site. Saudi Arabia, the UAE, India, Indonesia, Malaysia, Pakistan, China and others impose administrative fines or imprisonment. Jurisdictional compliance is the visitor's responsibility under our terms of use.
Can I be charged for AI-generated content depicting minors?
Yes, in every major jurisdiction. US 18 USC § 1466A and § 2256, UK Protection of Children Act 1978, EU CSA Directive 2011/93/EU (May 2024 amendment), Canada Criminal Code § 163.1, Australia OSA all reach AI-generated material. There is no jurisdiction where this is lawful.
What about marriage to an AI?
No US state, EU member state, or country recognises marriage to an AI. Marriage statutes universally require two human parties with legal capacity. Operator symbolic ceremonies inside apps and the 2018 Akihiko Kondo Hatsune Miku ceremony are not civil marriage. Our forward-looking FAQ walks the question in depth.
How do I find a lawyer for AI-companion-related questions?
US: state bar referral, Martindale-Hubbell, Avvo, ABA directory. EU: e-Justice Portal. UK: Law Society's Find a Solicitor. Look for technology, AI, consumer-protection, and First Amendment specialties. The EFF and ACLU maintain directories of cooperating attorneys for First Amendment matters.
Related reading
Parent silo: Are AI Companions Safe?, the umbrella topic-pillar covering safety beyond legal (privacy, mental health, minor protection, legal).
Sibling topic-pillars: AI Companion Privacy & Data • AI Companion Mental Health • AI Companions and Minors.
Down the silo:
- Is AI girlfriend illegal? State-by-state, granular per-state statute walkthrough.
- AI CSAM laws explained, deep-dive on § 1466A, § 2256, Protection of Children Act, EU CSA Directive.
- Will AI girlfriends become legal to marry?, legislative outlook and cultural context.
- Is AI girlfriend haram?, religious-jurisprudence treatment for Muslim-majority readers.
Operator practice: Editorial methodology • Affiliate disclosure • Errata log.
Last verified 2026 · See errata log for any post-publish corrections · Editor: Alexandra Joly · Methodology · Editorial process · Affiliate disclosure
Try our compliance-clean pick, Candy.ai →