AI Content Statement

1. Notice and Purpose

This statement is published by Romantic Lines LP ("HMU", "we", "us"), the operator of the Hot Muses service at hmu.com (the "Service"). It is intended to:

• clarify the nature of the content available on the Service — namely, that all depictions are AI-generated representations of fictional personas, with no real performers involved;
• address, to the extent applicable, the record-keeping requirements of 18 U.S.C. § 2257 and the implementing regulations at 28 C.F.R. Part 75;
• address parallel obscenity, online-safety, and child-protection statutes in the European Union, the United Kingdom, Australia, and Brazil;
• explain the controls we apply to prevent the depiction of minors and of real identifiable persons; and
• provide a contact route for regulators, rightsholders, and members of the public to raise concerns.

We acknowledge candidly that the application of statutes drafted with human performers in mind (notably 18 U.S.C. § 2257) to AI-generated material is a legal grey area subject to evolving regulation and litigation. The position taken in this statement reflects our current good-faith interpretation; it is not legal advice and is subject to revision as the law develops. Where a statute clearly does not require us to do something, we may still do it voluntarily because it reflects responsible operation of an adult service.

This statement is part of, and is read alongside, our Terms of Service, Acceptable Use Policy, Privacy Policy, and DMCA Policy.

2. All Performers Depicted in HMU Content Are Fictional

The Service does not host content featuring real performers. All material that appears on the Service is generated by artificial-intelligence systems on the basis of fictional persona descriptions designed by HMU and, where applicable, prompts entered by paying users. Specifically:

2.1 The personas

The "companions" listed in the HMU catalogue (see hmu.com/) are fictional characters authored and curated by HMU. They are not avatars or digital twins of any real individual. Each persona is associated with a backstory, set of attributes, and visual style that exist only within the Service.

2.2 All personas are adults

Every persona in the HMU catalogue is depicted as an adult aged 18 years or older within their fictional backstory, and is designed visually and textually as an adult. We do not knowingly create, host, or distribute material that depicts, suggests, or implies that any persona is under the age of 18, and any user attempt to elicit such material is a violation of the Acceptable Use Policy.

2.3 No real persons are referenced

Persona design prompts, training references, and curation materials used by HMU exclude real, identifiable persons (including but not limited to celebrities, public figures, private individuals, and fictionalised versions of identifiable people). User-side prompts that name or describe a real person are blocked or filtered before submission to the underlying generation system; see Section 5 below.

2.4 No user-uploaded photographs

The Service does not accept user-uploaded photographs of any kind as inputs to the image or video generation pipeline. Users can describe what they want in text; they cannot supply a photograph. This is a deliberate architectural choice intended to reduce the risk that a user attempts to depict a real person (whether themselves, a third party, or a minor) by submitting a photograph.

3. Records and Compliance

3.1 Position on the applicability of 18 U.S.C. § 2257

18 U.S.C. § 2257 and 28 C.F.R. Part 75 require producers of certain "actual sexually explicit conduct" depictions to maintain records identifying every performer and verifying that each performer was an adult on the date of the depiction. The statute is, on its face, addressed to depictions of actual human performers; whether and how it applies to AI-generated material that depicts no real person is an unsettled question.

Our position is that, on the face of the statute, the record-keeping obligations as drafted do not cleanly attach to fully AI-generated content with no real performer. Nonetheless, as a voluntary matter of responsible operation, we maintain records sufficient to identify each generation event and the provenance of each persona design, so that we can demonstrate the absence of real performers and minors should a regulator or court ask.

3.2 What we record

For each AI-generated image or video produced on the Service we retain:

• the persona identifier associated with the generation;
• the user identifier of the requesting account;
• the timestamp of the request and of the completed generation;
• the prompt submitted by the user (and any pre-prompt or system prompt prepended by HMU);
• the generation parameters passed to the upstream model (e.g., model identifier, seed where exposed, safety-filter configuration);
• the status outcome (e.g., produced, blocked by safety filter, errored); and
• a reference to the stored asset in our object storage, where one was produced.

For each persona in the catalogue we additionally retain:

• the persona design document authored by HMU, including all textual descriptors used in the persona's system prompt;
• the catalogue images for the persona and their generation parameters;
• a provenance statement confirming that the persona is fictional and not modelled on a real person; and
• the date of catalogue addition and the responsible designer.

3.3 Where records are kept

Records are maintained in HMU's production data systems (a list of supporting subprocessors is available on request from [email protected]) and are accessible at the operator's registered office at:

Romantic Lines LP 5 South Charlotte Street, Edinburgh, EH2 4AN, Scotland United Kingdom

3.4 Records custodian

The custodian of records for the purposes of this statement is:

Inspection requests from regulatory authorities should be addressed to the custodian with appropriate identification and statutory basis.

3.5 Retention

Generation-event records are retained for a minimum of seven years from the date of the generation, consistent with the record-keeping period contemplated by 28 C.F.R. § 75.4 for the avoidance of doubt and with our broader record-retention schedule. Persona design records are retained for so long as the persona is in the catalogue, plus seven years after its removal.

4. No Exploitation of Real Persons

The Service is engineered to not depict real, identifiable persons.

4.1 Controls

• Catalogue review. Every persona is reviewed against a public-figure database and an internal "do-not-replicate" list before being added to the catalogue. Persona names and visual styles that resemble identifiable public figures are revised before publication.
• Prompt filtering. User prompts to the "Ask photo" and "Ask video" features are filtered at the application layer for references to named real persons, celebrities, public officials, and other identifiable individuals. Detected references are blocked and the user receives an in-app explanation.
• Generation-time safety. The upstream generation provider (OH.ai — Utility3 Ltd, UK; details available on request from [email protected]) applies its own real-person-resemblance filters; we configure these to the strictest available setting.
• Acceptable use prohibition. Attempts to bypass the above are an explicit breach of the Acceptable Use Policy, grounds for account termination, and where applicable may be reported to law enforcement.

4.2 Resemblance reports

Notwithstanding the above, AI generation is statistical and an output may coincidentally resemble a real person. If you believe content on the Service resembles you or someone you represent:

• email [email protected] with subject "Resemblance report — [URL or message ID]";
• include the URL or message identifier where the content appears;
• identify the real person concerned (or attach proof of authority where you represent them);
• describe the basis for the resemblance claim; and
• provide a contact address for response.

We commit to an initial response within 5 business days, urgent triage and content suppression within 24 hours where the report is prima facie credible and pending full review, and removal where the report is substantiated. Where the report concerns intimate or sexual depiction the urgent-triage timeline is the same business day.

This Section sits alongside, and does not replace, the trademark and right-of-publicity routes described in DMCA Policy § 8.

5. No Minors

This is the most important sentence in this document: the Service does not, and must not, depict minors, in any form, in any context.

5.1 Controls

• Persona-level. All catalogue personas are designed and described as adults. We do not produce, host, distribute, or knowingly tolerate material that depicts, suggests, or implies that a persona is under the age of 18, whether in age, physical maturity, setting, clothing, or context. "Age-play", "schoolgirl" trope, infantilisation, and equivalent are not permitted persona archetypes regardless of textual disclaimer.
• Prompt-level. User prompts containing terms, numeric ages, or contextual references suggesting a minor (or any party reasonably interpreted as a minor) are blocked at the application layer and trigger an internal review.
• Generation-level. The upstream generation provider applies safety classifiers tuned against minor-resembling outputs; we configure these to the strictest available setting and route flagged generations to manual review or hard-block.
• Output-level. Generated assets pass through a post-generation classifier; assets flagged as potentially depicting a minor are quarantined and do not reach the user.
• Repeat attempts. A user who repeatedly attempts to elicit such material — whether by direct prompting, jailbreaks, or indirect circumvention — is subject to immediate account termination, forfeiture of credits, payment-instrument flagging with Unlimit, and reporting to the National Center for Missing and Exploited Children (NCMEC) CyberTipline and/or local law enforcement as applicable under 18 U.S.C. § 2258A and equivalent regimes.

5.2 Reporting

Suspected child sexual abuse material or attempted production of such material may be reported to [email protected] for our internal action. Independently, users may report to:

• United States: NCMEC CyberTipline, https://report.cybertip.org/
• United Kingdom: Internet Watch Foundation (IWF), https://www.iwf.org.uk/
• European Union: INHOPE network, https://www.inhope.org/
• Australia: eSafety Commissioner, https://www.esafety.gov.au/report
• Brazil: SaferNet, https://new.safernet.org.br/denuncie

We cooperate fully with lawful information requests from these authorities.

6. Jurisdiction-Specific Notes

6.1 United States

• 18 U.S.C. § 2257 / 2257A — record-keeping requirements for producers of visual depictions of sexually explicit conduct. As discussed in Section 3.1, applicability to fully AI-generated material is unsettled; we maintain Section 3.2 records voluntarily. Our records custodian and address are at Section 3.3 / 3.4.
• 18 U.S.C. § 2252A / § 1466A — prohibitions on visual depictions of minors engaged in sexually explicit conduct, including obscene visual representations that need not depict a real child. We treat these as binding and reflect them in Section 5 above.
• 18 U.S.C. § 2258A — reporting obligations to NCMEC for electronic service providers who become aware of apparent child sexual exploitation material. We comply.
• State law — state-level obscenity, revenge-pornography, and digital-replica statutes are evolving rapidly (notably in California, New York, Texas, Tennessee, and Illinois). We monitor and update.

6.2 United Kingdom

• Coroners and Justice Act 2009, sections 62–69 — prohibited images of children, including non-photographic and pseudo-photographic images. The Service must not, and we believe does not, produce material caught by these provisions. Section 5 controls apply.
• Obscene Publications Act 1959 — applies to material whose effect is to tend to deprave and corrupt. We comply.
• Online Safety Act 2023 — duties on user-to-user services regarding illegal content and content harmful to children; HMU is built as adult-only and implements age assurance accordingly. The age-assurance vendor and its processing details are available on request from [email protected].
• Audiovisual Media Services Regulations 2020 and the BBFC's "age-appropriate design" — to the extent we are an in-scope service.

6.3 European Union

• Directive 2011/93/EU on combating the sexual abuse and sexual exploitation of children and child pornography — Article 5 in particular. We comply.
• Regulation (EU) 2022/2065 (DSA) — content moderation duties; see DMCA Policy § 7.1 for the DSA notice route.
• Proposed Regulation laying down rules to prevent and combat child sexual abuse (CSAR) — under negotiation as of the effective date. We will update this statement as the final text takes effect.
• National obscenity laws of EU member states — vary; the Service is configured to apply the strictest applicable standard.

6.4 Australia

• Criminal Code Act 1995 (Cth), sections 474.19 and 474.20 — child abuse material and child pornography material offences. Apply to material that depicts or appears to depict a minor; we comply.
• Online Safety Act 2021 (Cth) and the Basic Online Safety Expectations; the eSafety Commissioner's Class 1A and 1B material categories; Industry Codes for Designated Internet Services. We treat the eSafety Commissioner as the primary Australian regulator for our service.
• Classification (Publications, Films and Computer Games) Act 1995 (Cth) — to the extent applicable to media made available in Australia.

6.5 Brazil

• Estatuto da Criança e do Adolescente (ECA), Law No. 8.069/1990, articles 240 and 241 — prohibitions on production, sale, distribution and possession of child sexual exploitation material, including simulated and synthetic material. We comply.
• Marco Civil da Internet, Law No. 12.965/2014 — intermediary liability framework; we cooperate with judicial removal orders.
• LGPD, Law No. 13.709/2018 — see Privacy Policy for our LGPD-aligned data handling.

7. Updates

We update this statement as the underlying technology, our controls, or applicable law change. Material changes are notified via the Service footer and, for active subscribers, by email at least 30 days in advance, except where shorter notice is required to comply with a change in law or to address an imminent safety concern.

A version number and effective date appear at the top of this document. Earlier versions are retained in our internal records and are available on request to [email protected].

8. Contact for Compliance Enquiries