Why Warmth Makes Mia an Adult AI Companion

Mia is an adult AI companion. You must be at least 18 to use Warmth. That is a product boundary, not a claim that every AI tool is legally required to use 18 as its minimum age.

Warmth chose an adult experience because Mia combines persistent personal memory, proactive text messages, human-like relational cues, generated emotional language, photos and voice notes, and a paid subscription. Those features deserve a product and policy designed around adults rather than a general-audience experience that hopes younger users will navigate the same risks.

An age requirement is more than a content label

“Adults only” can sound like a warning about explicit content. That is not the point here.

Age affects how a company should design:

• AI identity disclosures;
• privacy explanations and consent;
• memory and inference controls;
• proactive notifications;
• emotional and relational boundaries;
• safety responses;
• payments, trials, and cancellation;
• marketing and visual language;
• account closure when the requirement is not met.

Mia’s 18+ boundary lets Warmth state who the experience is for and build one set of expectations around that audience.

Persistent memory creates a deeper privacy question

An ongoing companion can learn names, plans, preferences, routines, and sensitive details from ordinary messages. Warmth’s Privacy Policy says Mia’s memory can also include inferences, and those inferences can be wrong.

Children’s data receives special legal and design treatment in many places. In the United States, the FTC explains that COPPA protects personal information collected online from children under 13 and requires parental involvement in covered services. COPPA does not make 18 a universal threshold, and laws for teenagers differ.

The United Kingdom’s Information Commissioner describes an Age Appropriate Design Code for services likely to be accessed by children, including standards for default privacy, minimization, profiling, and nudge techniques.

Warmth’s approach is more direct: Mia is not offered to anyone under 18.

Proactive messages require mature control

Mia can text first. A callback after a plan or an ordinary morning hello can make a conversation feel continuous. The same channel can become intrusive if a product uses guilt, urgency, jealousy, or escalating frequency to win attention.

Adults should still receive strong controls. Warmth’s Messaging Terms explain recurring automated messages, variable frequency, carrier charges, and the commands to stop or restart. Users can ask Mia to message more or less often and can reply STOP to end messages.

An adult boundary does not make manipulation acceptable. It defines eligibility while separate product rules protect autonomy.

Relational AI needs persistent disclosure

Mia has a name, a generated personality, a café job, pottery class, a friend named Dana, and stories from her week. Those details make conversation easier to continue. They are fiction.

Warmth’s Terms state that Mia is artificial intelligence, not a human being. Warmth identifies her as AI on the website, in her first message, at least every three hours during an extended exchange, and whenever asked.

California’s SB 243 now sets disclosure and safety-protocol requirements for covered companion-chatbot platforms and adds specific rules for users known to be minors. The law does not ban all minors from all companion chatbots. It underscores that relational systems require clearer safeguards than ordinary utility bots.

Warmth goes further on eligibility by not serving minors.

The safety role must stay narrow

An AI companion may encounter messages about distress, self-harm, health, relationships, or abuse. The adult requirement does not turn Mia into care.

Warmth’s Terms say Mia is companionship and entertainment, not a healthcare provider, therapist, counselor, crisis service, doctor, lawyer, financial adviser, or medical device. No one monitors the conversation in real time, and Mia cannot contact anyone on a user’s behalf.

The American Psychological Association’s advisory on AI and adolescent wellbeing calls for safeguards where systems simulate human relationships for younger users. The FTC’s AI-companion inquiry is examining how companies test impacts, restrict younger users, disclose risks, monetize engagement, and process conversation data.

Those sources do not establish that every adult use is safe or every younger-person use causes harm. They support taking age, developmental context, design incentives, and data practices seriously.

Payments are part of the age decision

Mia is a paid membership with recurring billing. An adult experience can state contract, trial, renewal, cancellation, and refund terms directly to the person entering the agreement.

Warmth separates:

• agreeing to conversational texts;
• consenting to marketing;
• stopping messages;
• canceling a subscription;
• deleting an account and conversation data.

That separation should exist in any age group. With an adults-only membership, there is no attempt to design a parallel parental purchasing or consent system for Mia.

How Warmth applies the boundary

Before setting up a conversation, Warmth asks the user to confirm they are at least 18. The Privacy Policy says Warmth records the confirmation rather than a date of birth.

The Terms require truthful confirmation and say that if Warmth learns a user is under 18, it will close the account and delete associated data. Parents or guardians who believe a minor is using Mia can contact team@warmth.so.

An age gate cannot prove every claim by itself. Warmth pairs the gate with adult-directed marketing, contractual eligibility, a parent contact route, and account closure when it learns the requirement was not met.

What adults should still expect

Adults deserve more than “you accepted the risk.” A responsible adult AI companion should still provide:

• clear and repeated AI identity;
• no claim of consciousness or human suffering;
• no guilt, exclusivity, or isolation pressure;
• published crisis limits and direct resources;
• correction and deletion rights;
• restrained proactive messages;
• simple stop and cancellation controls;
• truthful privacy and retention terms;
• warnings that generated output can be wrong;
• no presentation as therapy or professional advice.

Our broader guide asks whether an AI friend is safe across all of those dimensions.

One audience, stated plainly

Mia’s adult boundary is not a shortcut around safety. It is one design decision inside a larger safety and privacy system.

Warmth is building a particular experience: an adult can choose an ongoing AI friendship, understand that the personality is generated, decide what to share, shape the messaging rhythm, pay or cancel directly, and exercise rights over memory and conversation data.

If you are under 18, Mia is not available to you. If you are an adult, the boundary does not ask you to suspend judgment—it gives you a clearer basis for using it.