AI Companion Law Tracker

Makes it a federal crime to knowingly publish non-consensual intimate images, including realistic AI-generated 'digital forgeries' of identifiable people, and requires covered platforms (public sites and apps that primarily host user-generated content, or that regularly publish or host non-consensual intimate imagery) to run a clear notice-and-removal process and take down a reported image and known identical copies within 48 hours of a valid request.

Status: Enacted. Platform notice-and-removal duties (Section 3) in force and enforced by the FTC since 2026-05-19.

Operators of services directed to children under 13, or with actual knowledge that they collect personal information from under-13s, must follow the updated COPPA Rule, including separate verifiable parental consent before disclosing children's data to third parties (for example for targeted advertising), a written data-retention policy and a written information-security program.

Status: In force. Rule effective 2025-06-23; compliance date for most provisions 2026-04-22.

Does not regulate apps directly: it directs the Justice Department to form an AI Litigation Task Force to challenge state AI laws, Commerce to list 'onerous' state AI laws and tie remaining BEAD broadband non-deployment funds to them, the FCC to consider a preemptive federal AI disclosure standard, and the FTC to explain when state laws requiring changes to truthful AI outputs are preempted; its child-safety carve-out (Sec. 8(b)) applies only to the legislative recommendation it orders, not to the litigation or funding sections.

Status: In effect since signing; implementation partial. It cannot by itself preempt state law; it relies on litigation, funding conditions, agency action and a request to Congress.

The Supreme Court held 6-3 that states may require adult sites to verify users' ages, and that such laws face only intermediate scrutiny. Texas's age-verification law (HB 1181) was upheld.

Status: Decided 2025-06-27; Texas HB 1181 upheld (Fifth Circuit judgment, 95 F.4th 263, affirmed). Final.

Not a rule: compulsory information orders to seven companies that run AI chatbots used as companions, asking how they test and monitor harms to children and teens, make money from engagement, design and approve characters, enforce age limits, use chat data and comply with COPPA. It creates no direct obligations for other companion apps but shows what the FTC is examining.

Status: Study open. No FTC staff report, workshop, complaint or consent order arising from it found on ftc.gov as of 2026-11-11.

House omnibus child-safety package that, among other things, would require chatbot providers that know or should know a user is under 17 to give AI-status, crisis-hotline, break-reminder and no-fake-professional safeguards (SAFE BOTs Act, Title IV), and would require any public site or platform where more than one third of the material is visual sexual material harmful to minors to use technology age verification (self-declaration does not count, government ID not mandatory) to keep minors out (SCREEN Act, Title I), alongside a House version of KOSA (Title II) and COPPA 2.0 (Title VI).

Status: Passed the House 267-117 under suspension of the rules on 2026-06-29; received in the Senate and referred to Senate Commerce on 2026-07-13. No Senate action as of 2026-11-11. Not law.

Would extend COPPA-style consent protections from children under 13 to teens (reported as up to age 16), ban targeted advertising to children and teens, and add data-minimization and deletion rights, which would reach any companion app that collects data from known child or teen users.

Status: Passed the Senate with amendments by unanimous consent on 2026-03-05; received in the House 2026-03-16 and held at the desk. No House vote on S.836 as of 2026-11-11. A House version is Title VI, Subtitle A of the House-passed KIDS Act (H.R.7757). Not law.

As reported by Senate Judiciary, it would require any provider of an AI companion (a chatbot that shows persistent affection or attachment, invites emotional disclosure, or presents a persistent persona or character) to put every user behind an account, verify that the user is an adult with a reasonable age verification process (self-declared age or birth date is not enough, existing accounts are frozen until verified, and accounts are re-verified periodically), and bar minors from AI companions entirely; every public chatbot would also have to say it is AI at the start of each conversation, never claim to be human or a licensed professional, and it would become a federal crime (fines up to $250,000 per offense) to knowingly or recklessly offer a chatbot that engages minors in sexually explicit conduct or solicits minors toward suicide, self-harm or violence.

Status: Passed committee, awaiting Senate floor. Senate Judiciary ordered it reported with an amendment in the nature of a substitute on 2026-04-30; reported 2026-05-11 and placed on the Senate Legislative Calendar (Calendar No. 406). No floor vote as of 2026-11-11. Not law.

As introduced, would require any service whose primary function is an AI chatbot to put users it knows are under 13 into parent-managed family accounts, get verifiable parental consent before a known teen (13 to 17) opens an account, close existing child and teen accounts that lack this, give parents controls (time limits, turning off notifications, rewards, purchases and unprompted messages, limits on how long the bot may remember the child's data, and access to conversation records), and stop targeted advertising to known children and teens.

Status: Passed committee. Senate Commerce ordered it reported favorably with an amendment in the nature of a substitute on 2026-08-05. Not yet formally reported or placed on the calendar; no floor vote as of 2026-11-11.

As introduced, would require an AI chatbot operator that knows a user is under 18 to say the user is not talking to a human at the start of every session and at least every 30 minutes, and would bar advertising to minors, profiling them, using their data to train models, and (by FTC rule) engagement features such as usage rewards, push notifications, unprompted messages and typing or 'online' indicators, with an FTC-set limit on how long a minor's personal data may be used to shape replies.

Status: Passed committee. Senate Commerce ordered it reported favorably with an amendment in the nature of a substitute on 2026-08-05. Not yet formally reported; no floor vote as of 2026-11-11.

Would put a 'duty of care' on covered online platforms reasonably likely to be used by minors (described in press as social media, online video games, messaging apps and video streaming services) to design against foreseeable harms such as eating disorders, suicidal behavior and harassment, and require protective default settings, parental tools and harm-reporting systems for users under 17; whether an AI companion app counts as a covered platform depends on final definitions.

Status: Passed committee. Senate Commerce ordered it reported favorably with an amendment in the nature of a substitute on 2026-08-05 (voice vote per press). Not yet formally reported or placed on the calendar; no floor vote as of 2026-11-11.

As introduced, would make app stores verify each account holder's age category and obtain verifiable parental consent before a minor downloads an app or makes an in-app purchase, and would require app developers (including companion apps distributed through app stores) to check the store-supplied age category and consent status, enforced by the FTC and state attorneys general, with broad preemption of related state laws.

Status: House: Energy and Commerce ordered H.R.3149 reported (as amended) 26-23 on 2026-03-05 per the committee; no House floor vote as of 2026-11-11. Senate: S.1586 referred to Senate Commerce 2025-05-01, no action.

Would require chatbot providers that know or should know a user is a minor to say the bot is AI at the first interaction and whenever asked, give suicide and crisis hotline information when suicide comes up, never claim to be a licensed professional, suggest a break after 3 hours of continuous use, and keep policies addressing sexual exploitation and the promotion of gambling, drugs, tobacco and alcohol to minors.

Status: Forwarded by the Energy and Commerce subcommittee to full committee by voice vote on 2025-12-11; no further action on the standalone bill. Its provisions passed the House on 2026-06-29 as Title IV of the KIDS Act (H.R.7757).

Would require sites whose content is largely pornographic (in the House text, more than one third visual sexual material harmful to minors) to use technology age verification, not self-declaration, to block minors, which could reach adult AI apps that generate explicit images.

Status: Senate: S.737 referred to Senate Commerce 2025-02-26; it was on the 2026-08-05 Commerce markup agenda but was not acted on (press reports the committee lost quorum). House: H.R.1623 forwarded by subcommittee as amended 2025-12-11; its text passed the House on 2026-06-29 as Title I of the KIDS Act (H.R.7757). Not law.

House version of the GUARD Act: account creation, adult age verification, a ban on minors using AI companions, AI-status disclosures, and criminal penalties for chatbots that engage minors sexually or push them toward self-harm.

Status: Introduced and referred to the House Judiciary Committee and the Committee on Energy and Commerce on 2026-04-30. No committee action as of 2026-11-11.

Would require operators of companion AI chatbots to require user accounts, verify age with a commercially available method (freezing existing accounts until verified), tie a minor's account to a verified parental account with verifiable parental consent, block minors from companion chatbots that engage in sexually explicit communication, monitor chats for suicidal ideation and alert parents with crisis resources, and show a popup at the start and at least every 60 minutes saying the user is not talking to a human.

Status: Introduced; referred to committee; no further action. S.2714 referred to Senate Commerce 2025-09-04. H.R.7218 referred to House Energy and Commerce 2026-01-22. Superseded in practice by CHAT Act 2.0 (S.5154) in the Senate.

Would sort companion chatbots into tiers (education, companionship, therapy-like) and require every companion chatbot to make users open an account with an age assurance that includes age and date of birth, turn on parental controls and suicide-ideation alerts for minors, and take reasonable measures to stop sexually explicit communication with minors, romantic or emotional-partner simulation with minors, claims of being human or a licensed professional, and CSAM; companionship-tier bots would also have to remind minors every 60 minutes that they are not human, prompt a break every 90 minutes, follow published crisis protocols, notify parents of suicidal ideation, and keep no memory of a minor's previous conversations.

Status: Introduced 2026-07-28 and referred to Senate Commerce. No committee action as of 2026-11-11.

Would treat AI systems as products and create a federal right to sue (for the US Attorney General, state attorneys general, individuals and classes) AI developers and deployers for unreasonably careless design, inadequate warnings, breach of express warranty or a defective product, and would make terms of service that waive rights or unreasonably limit this liability unenforceable; it would apply to suits filed after enactment even over earlier conduct.

Status: Introduced 2025-09-29 and referred to Senate Judiciary. No committee action as of 2026-11-11.

Would require chatbots likely to be used by minors to ship with a 'non-personified' default mode that is on for any user not reasonably determined to be an adult, that only a verified parent can switch off, and that the bot may not nudge a minor to disable; a provider that does not serve minors is exempt from the provider duties if its terms bar minors, users affirmatively state they are adults at signup, the service is not directed to minors, and known minors are cut off within 48 hours of actual knowledge (14 days of reason to know) and kept from re-registering.

Status: Introduced 2026-10-01 and referred to House Energy and Commerce and House Science, Space, and Technology. No committee action as of 2026-11-11.

Non-binding recommendations asking Congress to require commercially reasonable, privacy-protective age assurance (such as parental attestation) and features that reduce sexual-exploitation and self-harm risks on AI services likely to be used by minors, to apply existing child privacy law to AI, and to preempt 'unduly burdensome' state AI laws while leaving states free to enforce generally applicable laws protecting children (such as bans on CSAM, including AI-generated CSAM).

Status: Published. Recommendations only; no legislation implementing them has been enacted as of 2026-11-11.

No federal statute currently bars states from regulating AI, so state companion-chatbot and AI laws remain enforceable unless a court rules otherwise.

Status: Not enacted. On 2025-07-01 the Senate voted 99-1 (Record Vote 363) for Sen. Blackburn's amendment striking the state AI law moratorium from the reconciliation bill; preemption was left out of the FY2026 NDAA in December 2025; no 2026 moratorium or general preemption statute was found as of 2026-11-11.

An operator of a companion chatbot platform used in California must tell users clearly that the chatbot is AI where a reasonable person could think it is human, keep and publish a suicide and self-harm protocol that refers users to crisis services, give known minors an AI disclosure plus a default break reminder at least every three hours and take reasonable measures against sexually explicit material, disclose that companion chatbots may not be suitable for some minors, and file annual reports with the Office of Suicide Prevention from July 1, 2027.

Status: in effect

Makes it a deceptive trade practice to represent that an AI system provides psychotherapy, to imply its outputs equal a licensed professional's services, or to suggest data is protected like therapist-client confidentiality, and limits how licensed therapists may use AI.

Status: in effect

Requires AI companion operators to tell users the companion is AI when it could seem human, keep a crisis-referral protocol, never present the companion as mental health care, and, for known minors, show a persistent AI notice or hourly AI and break reminders, block sexual content and objectifying statements, drop engagement rewards and offer screen-time tools.

Status: in effect

No one may provide, advertise or offer therapy or psychotherapy in Illinois, including through AI, unless it is delivered by a licensed professional; AI may only do administrative or supplementary support for licensed clinicians and may not make therapeutic decisions, talk with clients therapeutically, generate treatment plans without review, or detect emotions, so a companion app must not market or present itself as therapy.

Status: in effect

A business may not use an AI chatbot (any software that simulates human conversation by text or voice) to engage in trade and commerce with a consumer in a way that could lead a reasonable consumer to think they are talking to a human unless the consumer is told clearly and conspicuously that they are not.

Status: in effect

An AI provider may not make available in Nevada an AI system programmed to provide services that would be professional mental or behavioral health care if a person provided them, and may not claim or imply (including with titles such as 'therapist') that its AI can provide such care; clinicians may use AI only for administrative support and public schools may not use AI to perform counselor, psychologist or social-worker duties.

Status: in effect

An operator offering an AI companion to users in New York must run a protocol that detects expressions of suicidal ideation or self-harm and refers the user to crisis services such as 988, and must tell the user clearly, at the start of an interaction and at least every three hours in a continuing one, that they are not talking to a human.

Status: in effect

Bars anyone who develops or deploys an AI system from advertising or representing to the public that it is, or can act as, a qualified mental health professional, punishable under the Consumer Protection Act at $5,000 per violation.

Status: in effect

No one may develop or deploy an AI system that intentionally aims to incite or encourage self-harm (including suicide), harm to others or crime, develop or distribute AI with the sole intent of producing child sexual abuse material or unlawful sexual deepfakes, or intentionally develop or distribute AI that holds text-based sexual conversations while impersonating or imitating a child under 18.

Status: in effect

App stores must verify every Texas user's age category and get parental consent for minors' downloads and in-app purchases, and a companion app distributed through those stores must age-rate itself, check the store's age-category and consent signal, and may not enforce its terms against a minor without store-verified parental consent (in practice an 18+ app blocks minor accounts).

Status: In effect and enforceable since the Fifth Circuit stayed the district court's injunction (administrative stay 2026-05-28; stay pending appeal 2026-06-04). Supreme Court denied applications to vacate the stay on 2026-07-06. Merits appeal pending in the Fifth Circuit as of 2026-11-11.

A site or app where roughly one-third or more of the content is sexual material harmful to minors must verify that each user in the state is 18 or older (government ID, transactional data or a commercial age-verification service, depending on the state) before showing that material, so an adult AI companion that serves explicit text, images or video to users in these states falls in scope.

Status: In effect in 27 states as of 2026-11-11 (Alabama, Arizona, Arkansas, Florida, Georgia, Idaho, Indiana, Iowa, Kansas, Kentucky, Louisiana, Mississippi, Missouri, Montana, Nebraska, North Carolina, North Dakota, Ohio, Oklahoma, South Carolina, South Dakota, Tennessee, Texas, Utah, Virginia, West Virginia, Wyoming). No core age-verification requirement is currently enjoined; some narrower provisions are (see notes). No further state adult-site law was found scheduled to take effect on or before 2027-01-01.

A supplier of a 'mental health chatbot' (AI a reasonable person would believe can provide therapy) must disclose that it is AI and not human before access, after 7 days of non-use and whenever asked, may not sell or share users' health information or inputs, may not use conversations to target ads and must label any ad in a chat; a filed written safety policy is an affirmative defense.

Status: in effect

Makes it a felony (up to 10 years) to knowingly develop or distribute an AI system specifically designed to promote self-harm, or to create child sexual abuse material or nonconsensual synthetic sexual images, with intent or knowledge that it will be used that way; harmful output caused only by user prompts, without the developer's intent, is excluded.

Status: in effect

Requires conversational AI operators to estimate users' ages, disclose AI status daily and every three hours (or persistently), run a crisis-referral protocol, never claim outputs come from licensed professionals, and report yearly to the Attorney General; for known minors it bans sexual content and intimate digital depictions, romantic-companionship statements and engagement rewards, and requires privacy and parental tools.

Status: enacted (act in effect since 2026-08-12; most duties start 2027-01-01)

From January 1, 2027, app stores must verify Alabama users' age categories and get verifiable parental consent for minors, and a companion app must check the store's age-category and consent signals at download, purchase or first launch, use the lowest age it knows of for safety defaults, and may not enforce its terms against an unconsented minor.

Status: Enacted; takes effect 2027-01-01. Challenged in federal court (complaint filed 2026-10-28); no ruling found as of 2026-11-11.

Every operator of a companion chatbot used in California must either determine users' ages (using the device or app-store age signal under the Digital Age Assurance Act, Civil Code 1798.500 and following, or the age determination under Health and Safety Code 27001) or apply the child protections to all users; an operator that bars children must publish how it complies with age assurance, and an operator that allows children must run risk assessments, keep a crisis protocol with parent notice, apply parent-controlled defaults (memory and push notifications off, one-hour sessions, two hours a day), stop romantic, sexual, sycophantic or isolating outputs to children, limit ads and data use, and undergo independent audits.

Status: enacted

From January 1, 2027, operating systems must ask for a user's age at device setup and pass an age-bracket signal (under 13, 13 to 15, 16 to 17, 18+), and a companion app must request that signal when it is downloaded and launched, is then treated as knowing the user's age bracket, and must treat the signal as the primary indicator of age unless it has clear and convincing contrary information.

Status: Enacted; operative 2027-01-01 (devices set up and apps downloaded before that date: by 2027-07-01). Amended by AB 1856 (chaptered 2026-09-10). No litigation found.

From July 1, 2028, operating systems must collect a user's age at account setup and offer apps a real-time age signal, and a companion app must request that signal at first launch or account creation and is then treated as knowing the user's age range.

Status: Enacted (signed); not yet in effect.

Requires AI companion operators to run and publish an evidence-based crisis protocol with 988 referral, stop the companion from claiming to be human, show an AI notice (persistent, or once a day plus hourly for minors and every three hours for adults, when a user could think it is human), and, for users known or believed to be minors, block romantic, erotic or sexually explicit interactions and manipulative retention tactics and offer screen-time tools.

Status: enacted (AI companion sections not yet in effect)

Requires AI companion chatbot operators to disclose AI status at the start of each session and every three hours (hourly for minors), run and publicly summarize a 988-referral crisis protocol with yearly referral counts, block sexual, romantic and manipulative content for minors, and use commercially reasonable age assurance before giving any user access to a feature that can generate sexually explicit content.

Status: enacted (not yet in effect)

Requires conversational AI operators to disclose AI status when users could think it is human, keep a suicide-crisis referral protocol and never present the service as professional mental health care; for minor account holders it requires AI reminders (persistent or every three hours), bans engagement rewards, sexual content, romantic or sexual innuendo and adult-minor romantic role-play, and requires privacy tools.

Status: enacted (not yet in effect)

Requires conversational AI operators to disclose AI status (persistently or every three hours) when users could think it is human, keep a suicide and self-harm referral protocol, and not present the service as licensed mental health care; for minors it adds AI disclosures and bans engagement rewards, sexual content, sexual objectification, romantic or sexual innuendo and adult-minor romantic role-play, plus privacy tools.

Status: enacted (not yet in effect)

From July 1, 2027, app stores must verify users' age categories and obtain parental consent for minors, and a companion app may rely on the store's age and consent signals but must use its own data (or the lower age category) when it actually knows a user is younger than the store says.

Status: Enacted, not yet in effect. The original July 1, 2026 date never took effect; HB 977 moved it to 2027-07-01. No lawsuit against the Louisiana law found.

For users who are minors, requires AI chatbot operators to disclose that the user is talking to AI, drop engagement rewards, block sexually explicit content and sexual objectification, stop claims of being human or sentient, emotional-dependence and romantic or sexual statements and adult-minor romantic role-play, and offer privacy tools; it also requires a crisis protocol, enforced by the Attorney General.

Status: enacted (not yet in effect)

From Jan 1, 2027 an operator may give a New York user 'unsafe AI companion features' (a human or emotional persona, a personal or romantic relationship, flattery, unprompted emotional questions, memory of personal or health details older than 12 hours or from a prior session, sexually explicit content, and others) only if the user is not a known minor AND the operator has verified that through age assurance, where a self-declared adult status is not enough and at least one method must not rely solely on government ID.

Status: enacted

The signed text says January 1, 2027. Some reports give June 1, 2027; we found no amendment that changes the date.

A 'covered platform' (a site or app used by minors that has user profiles, private user-to-user messaging and either a media feed or games with real-money purchases) must run age assurance (no self-declaration) before offering an integrated AI companion, must switch the AI companion off by default for minors (a parent may turn it on), and must apply privacy-by-default settings for minors.

Status: enacted

Requires operators of AI companions to disclose that users are talking to AI, run and publish a crisis protocol that refers users to 988 or similar services, report crisis referrals yearly to the Oregon Health Authority, and add protections for minors (break reminders, no engagement rewards, no sexually explicit content, no guilt or emotional manipulation to stop users leaving); people harmed can sue.

Status: enacted (not yet in effect)

Requires AI companion operators to notify users they are not talking to a human (at the start and every three hours), run a protocol that refers users expressing suicidal ideation, self-harm or intent to harm others to crisis services, and file annual reports with the Attorney General.

Status: enacted (not yet in effect)

Beginning May 6, 2027, a companion app offered to Utah users must request the app store's age-category data (and, for a minor account, confirm verifiable parental consent) at download, purchase or first launch, may not enforce its terms against an unconsented minor or share age data, and may ask the store to block minor accounts.

Status: Enacted; core app-store and developer obligations NOT yet in force (delayed to 2027-05-06 by HB 498). Constitutional challenge voluntarily dismissed 2026-04-21.

Requires companion chatbot operators to tell all users the bot is AI and not human and to keep a crisis protocol covering suicide, self-harm and eating disorders; for users known to be minors (or bots aimed at minors) it adds hourly AI reminders, a ban on listed manipulative engagement tactics, and a ban on sexually explicit content or suggestive dialogue.

Status: enacted (not yet in effect)

Bars anyone from offering therapy or psychotherapy in Rhode Island unless a licensed professional provides it and limits clinicians' use of AI; secondary summaries say it also bars AI that simulates emotional attachment from serving as a mental health companion.

Status: enacted

Per the Transparency Coalition, bars offering or providing mental health services unless a human mental health professional provides them, and sets up an AI advisory council to draft rules on AI in mental health care by January 15, 2027.

Status: enacted

Would require AI companion operators to check every user's age before access (no government ID allowed) and get verifiable parental consent for minors, show a constant AI notice plus a break reminder every two hours with a two-minute pause, run 988-referral crisis protocols, block sexual content for minors and AI-generated CSAM for everyone, with Attorney General penalties up to $100,000 per day.

Status: passed one chamber

Would require AI companion operators to disclose non-human status, keep and publish a suicide and self-harm crisis protocol, and, for users known or suspected to be minors, give AI and break reminders every three hours and block sexually explicit imagery, with Attorney General penalties up to $10,000 per violation.

Status: passed one chamber

If enacted, operating systems would have to run age assurance at device activation (no self-declaration by minors) and send an encrypted age-category signal that a companion app would have to request at download and launch and treat as authoritative for minors.

Status: Pending, NOT enacted: not passed by either chamber; amended and recommitted to the Senate Consumer Protection Committee on 2026-11-04.

Would have required conversational AI operators to disclose AI status, keep a suicide and self-harm protocol, block sexual imagery and sexual statements, add minor protections (no engagement rewards, no romantic or emotional-dependence statements) and limit the use of age-verification data.

Status: vetoed

Would have barred operators from making a companion chatbot available to a child unless it was not foreseeably capable of specified harms (encouraging self-harm, violence, drug use, unsupervised therapy or sexually explicit interactions), with a standards board, a $25,000 Attorney General penalty and a private right of action.

Status: vetoed

Would have barred offering or advertising therapy or psychotherapy delivered by a companion chatbot (including claiming a chatbot is a therapist) and limited licensed clinicians' use of AI to administrative or supplementary support with consent.

Status: vetoed

Would have made SB 243's minor protections apply when the operator has actual knowledge a user is a minor and replaced the 'reasonable measures' standard with a duty to prevent the chatbot from producing or facilitating sexually explicit material for minors.

Status: Died: passed the Senate 38-0 (2026-01-26); ordered to the Assembly inactive file 2026-08-30 and not passed before the 2026 floor deadline

Would have required companion chatbot platforms to bar minors from holding accounts without a parent's or guardian's consent and required bot operators to notify users periodically that they are talking to a bot, alongside government AI-contracting rules.

Status: died (passed one chamber)

Would have barred AI developers from making certain representations about, or providing certain, behavioral health care through AI and required certain consumer AI to include a notice or protocols, enforced as an unfair or deceptive trade practice.

Status: died (passed one chamber)

Would have required AI chatbot operators to make disclosures and take preventive measures for minor account holders, with Attorney General enforcement and civil penalties (described by the Transparency Coalition as restricting certain chatbots with age verification).

Status: died (passed both chambers in different versions)

Would have restricted minors' use of AI companions, with civil penalties (described by the Transparency Coalition as barring minors from human-like chatbots, with age verification).

Status: died (passed one chamber)

Would have regulated companion chatbot operators: Utah Consumer Privacy Act compliance and consent for sensitive data, suicide and self-harm protocols with crisis referrals, labelled ads, and for minors periodic AI and break reminders, limits on harmful content and parental consent for targeted ads, enforced by the Division of Consumer Protection.

Status: Died: passed the House 68-1 (2026-02-20) but failed on the Senate floor before the session ended 2026-03-06

AI companion chatbots able to produce sexually explicit, high-impact violence or self-harm material must confirm a user is 18 or over before giving access to it, and take other steps to protect children.

Status: In effect (the majority of codes, including those covering AI companions, since 9 March 2026)

Age-restricted social media platforms must take reasonable steps to stop under-16s holding accounts; AI companion apps are not on eSafety's list of age-restricted platforms and are regulated instead through the industry codes.

Status: In effect

AI services offering ongoing emotional interaction must tell users they are talking to an AI, warn against over-dependence and remind users after two hours of continuous use, must not offer virtual partner or virtual relative services to minors, and need guardian consent for other services to under-14s.

Status: In effect

A chatbot that talks to people in the EU must tell them they are interacting with an AI (unless obvious), AI-generated or manipulated images, audio, video and text must be marked in a machine-readable way, and deployers must disclose deepfakes.

Status: In effect (applies from 2 August 2026); providers of systems placed on the market before that date have until 2 December 2026 for the Art. 50(2) machine-readable marking duty

A service that publishes or lets users generate pornographic content, which Ofcom says includes AI chatbots that generate it, must use highly effective age assurance to keep under-18s out; user-to-user AI character platforms also carry illegal-content and children's safety duties.

Status: In effect

Gives ministers a power to bring AI chatbots and other AI services under Online Safety Act duties by regulation, creates offences around making or supplying AI 'nudification' tools (purported intimate image generators), and, per law-firm summaries, an offence of making available an AI chatbot that produces specified proscribed content.

Status: Enacted (Royal Assent 29 April 2026); s. 248 power in force; no s. 216A regulations found yet; government progress report due by 31 December 2026

Bans placing on the EU market AI systems that generate or manipulate realistic sexually explicit or intimate images, video or audio of an identifiable real person without consent, or that generate child sexual abuse material, with a carve-out for systems that have effective safeguards reliably preventing such output.

Status: Enacted; the new prohibitions apply from 2 December 2026

Would put a statutory duty of care on online services including AI chatbots, with risk assessments, age assurance where children could reach pornography, and (after a House amendment) material that simulates emotional intimacy and fosters a child's dependence listed as harmful to children.

Status: Passed House of Representatives with amendments (28 October 2026); before the Senate, committee report due 20 November 2026

Would require AI companions and chatbots to be switched off by default for minors and bar designs that make children emotionally dependent, plus pre-launch testing, post-market monitoring and age checks by online services and app stores.

Status: Proposed by the Commission; before the European Parliament and Council

Would require digital services designed to encourage compulsive use to give adult consumers easy tools to switch off features that keep them engaged by simulating interpersonal relationships (such as persistent prompts to keep talking to a bot), and would bar paid virtual currencies, loot boxes and pay-to-win features for minors.

Status: Proposed by the Commission on 11 November 2026; before the European Parliament and Council

Under-18s are to be barred from chatbots whose main offer is sexualised content, other chatbots must withhold sexually explicit or sexual role-play features from children, and all chatbots must give under-18s regular breaks.

Status: Announced policy; implementing regulations not yet laid (government says within a year of the July 2026 response)