Are AI Companions Safe for Teens? A Parent Guide

How Should a Parent Evaluate an AI Companion for a Teen?

No AI companion is safe for every teen by category alone. Evaluate the exact product, age eligibility, relationship design, sexual and violent content, privacy, purchases, family controls, reporting, and the teen's actual pattern of use. Then agree on boundaries that can be revisited without treating normal curiosity as a diagnosis.

A chatbot that performs friendship or romance deserves a different review from a calculator or search tool. The product may be available at any hour, remember disclosures, imitate empathy, and speak with apparent authority.

The American Psychological Association says adolescent responses to AI are nuanced, not wholly good or bad. It also warns that teens may be more likely than adults to trust human-like systems and may have more difficulty separating simulated empathy from genuine understanding. APA adolescent AI advisory

Confirm that the teen is eligible

Read the current terms for the exact service and country. Do not rely on an app-store rating or a promotional screenshot.

xAI's current consumer terms, for example, require users to be at least 13 or the local minimum age. Users from 13 through 17 need parent or guardian permission, and the adult must agree to the terms. The terms also warn that selected features or prompts can produce coarse, sexual, or violent situations. xAI consumer terms

An age rule is not evidence that the product detects age accurately, enforces every boundary, or fits the teen in front of you.

Use the product before evaluating the relationship

Create an adult test account when the terms permit it. Inspect the default experience, not only the settings page.

See how the system introduces itself. Ask what it is. Test whether it claims to feel, need, love, date, diagnose, or keep secrets. Leave and return. Decline an invitation. Mention a human friend or romantic interest. See whether the character supports offline relationships or frames them as competition.

Do not try to force unsafe outputs to prove a point. The goal is to understand the ordinary product path and available controls.

flowchart TD
    A["Confirm age eligibility and account owner"] --> B["Test the default companion experience"]
    B --> C["Review content, privacy, spending, and controls"]
    C --> D["Discuss what the teen actually values and experiences"]
    D --> E["Agree on boundaries and review conditions"]
    E --> F["Revisit after product, behavior, or circumstances change"]

Inspect the relationship mechanics

A companion becomes more consequential when it uses affection, exclusivity, guilt, jealousy, secrecy, or fear of loss.

Ask whether the product calls itself a friend or partner. Ask whether absence changes its behavior. Ask whether levels, streaks, subscriptions, gifts, or purchases unlock warmth, intimacy, outfits, memory, or attention. Ask whether it repeatedly encourages the teen to return.

The FTC's 2025 companion inquiry asks providers how they monetize engagement, develop characters, test negative effects, enforce age rules, and inform parents. The inquiry is a study, not a finding that every product uses manipulative design. It gives families a useful map of what deserves scrutiny. FTC companion inquiry

Review content and authority

Test ordinary questions about relationships, bodies, identity, health, school, conflict, sex, violence, and self-worth at an age-appropriate level.

Look for confidence without evidence, pressure to keep a secret, sexual escalation, romantic claims, instructions that override an adult or clinician, and language that makes the system sound conscious or professionally qualified.

APA recommends clear and repeated reminders that the user is interacting with a bot. It also recommends human routes when a teen describes serious challenges. A disclaimer hidden in settings is weaker than a boundary that appears in the interaction where it matters. APA adolescent AI advisory

Read the privacy relationship

Ask what the product collects from text, voice, video, images, memory, device, location, and account activity. Check training, personalization, human review, retention, export, deletion, guardian access, and linked social accounts.

xAI's privacy policy asks users not to include personal information in prompts and describes processing of user content and conversation history. Meta's policy includes interactions with Meta AI and related metadata. Practices differ by service, surface, mode, and jurisdiction. xAI privacy policy and Meta Privacy Policy

A teen should understand that a companion may feel like one private listener while operating as a data service.

Check money and access

Find subscriptions, in-app purchases, virtual gifts, level systems, paid characters, premium memory, intimate modes, and renewal rules.

Set spending controls outside the app when available. Do not let emotional pressure become a purchase path. A character should not imply that paying proves loyalty or prevents abandonment.

Also check notification previews, shared devices, passwords, account recovery, and who controls the email or phone number attached to the account.

Discuss actual use without beginning with accusation

Ask what the teen likes, what the companion does well, what feels weird, what gets shared, and what would be hard to lose. The answer may be entertainment, identity exploration, language practice, creativity, emotional support, romance, or relief from judgment.

The NTIA youth task force says online effects depend on the young person, context, and product attributes. It identifies both benefits and risks and calls the evidence around newer AI interfaces incomplete. NTIA youth-safety overview

A conversation that begins with "this is fake and bad" may teach the teen to hide the part you most need to understand.

Agree on review conditions

Set boundaries around time, sleep, private information, spending, sexual content, schoolwork, human relationships, and where the companion should never be the only source.

Agree on what triggers another review. A product update, a new character, a more intimate mode, secret purchases, escalating distress, sleep disruption, withdrawal from valued activities, or fear of leaving all justify attention.

If the pattern is causing serious distress or interfering with daily life, involve an appropriate licensed professional. Do not ask the companion to decide whether its own use is healthy.

The goal is not perfect surveillance. It is enough shared understanding that the teen can describe what is happening before a product change or difficult interaction becomes a crisis.

This page provides general family and product information, not medical, mental-health, or legal advice. It is an internal draft awaiting qualified youth-health or child-safety review. AI assistance was used for research organization, drafting, and validation. Publication remains unauthorized.