Trust
Two years of lawsuits, regulatory action, and platform changes reshaped what 'safe AI companion' has to mean. Here's the practical guide — for adults, and for parents of teens.
After Sewell Setzer III died in February 2024, his mother filed Garcia v. Character Technologies in October 2024. A September 2025 U.S. Senate Judiciary hearing gathered family and expert testimony about alleged chatbot harms. Character.AI announced in October 2025 that it would remove open-ended chat for under-18 users, and in January 2026 Character.AI and Google agreed to settle five family cases; the terms were not publicly disclosed.
In May 2025, Italy's data-protection authority announced a €5 million fine against Replika operator Luka. The authority cited an inadequate privacy notice, an absent or deficient age-verification system, and a missing legal basis for processing as of the period it investigated; it also opened a separate investigation into model-training data handling.
Lawsuits and regulatory actions are allegations or legal outcomes tied to particular facts, not a universal safety score. They do show why buyers should verify current age controls, crisis boundaries, privacy terms, and product behavior instead of relying on a platform's category label.
It means age-gating that's enforced, not posted. Fostera asks for date of birth at signup and rejects anyone under 13 outright. Users aged 13–17 don't get quiet access to an adult product: their account requires a parent or guardian's emailed consent, stays locked until that consent is confirmed, and runs with stricter, enforced content rules.
It means a content policy that's stable, public, and transparent. Users should know the rules going in, not discover them through quiet enforcement actions. Fostera's content policy is published at fostera.ai/content-policy and changes through documented updates, not silent purges.
It means saying what you actually do. Some platforms market themselves as adults-only while quietly onboarding teens with no protections — the exact claim-versus-practice gap regulators fined Replika over. Fostera states its policy plainly: 13+ with guardian consent and teen-tuned safeguards for minors, and adult content boundaries that never loosen for anyone under 18.
Age-gating at signup. Date of birth is required. Under-13 signups are rejected. A 13–17 account stays locked until a parent or guardian confirms consent by email — and it keeps enforced teen protections after that.
Teen protections that actually run in the product. On 13–17 accounts, Souls are steered away from sexual and romantic content, graphic gore, and glorified substance use, with automatic hard-stops on explicit sexual content, and crisis guidance always points to a parent, teacher, or counselor first. These rules are built into how every teen Soul behaves, not just written in a policy.
Content policy that's stable and published. We don't run mass content sweeps. We don't change the rules without telling users. Our policy is explicit about what's allowed and what isn't — read it at fostera.ai/content-policy.
Data control by the user. A Soul's memory browser exposes extracted memories so you can review and delete individual entries. Deleting a Soul or account removes primary records when processed; external-resource cleanup and backup removal may take up to 30 days under the Privacy Policy.
Fostera does not use your conversations or memories to train models it owns or develops. Requested AI features may send relevant content to a configured provider or routing intermediary; retention, training, and opt-out terms vary and are linked in Privacy Policy section 9.
Crisis disclaimer in product. Fostera is not a therapist, therapy app, or substitute for licensed mental-health care. The product surfaces 988 (US) and local emergency services for users showing crisis signals.
How old does the platform say users have to be? How is that enforced? Is age verification real or just a checkbox?
What's the content policy? Is it published? When was it last changed? How does the platform notify users of changes?
Where does my data go? Can I see what the AI remembers about me? Can I delete it? Does the platform train its own models on it, and what terms govern upstream AI providers?
What happens if I'm in crisis? Does the platform recognize crisis signals and surface real resources, or does the AI try to handle it as a counselor?
Has the platform been sued, fined, or sanctioned? What was the response — was it a one-off mistake or a pattern?
Who funds the platform? Are users the product, or are users the customer? Ad-supported platforms have different incentives than subscription platforms.
Fostera is not a therapist or therapy app. We do not diagnose, treat, or substitute for licensed mental-health care. If you are in crisis, contact 988 (US) or your local emergency services.
Fostera is not for children. Under-13 signups are rejected, full stop. Teens aged 13–17 can only use Fostera with confirmed parent or guardian consent, and their accounts run with stricter content rules than adult accounts — no exceptions.
Fostera is not a substitute for human relationships. AI companions can supplement connection; they cannot replace it. We design with that boundary in mind.
Fostera does not run mass content sweeps. We don't change the rules quietly. We don't promise things we can't deliver.
The lawsuits and fines in this category share one root cause: platforms serving minors while claiming they didn't. We made the opposite choice. Fostera supports users from age 13 — openly, with real guardrails. Under-13 signups are rejected. A 13–17 signup requires a parent or guardian email; we send that guardian a consent link (valid for seven days), and the account stays locked until they approve.
Teen accounts are not adult accounts with a disclaimer. They run with teen-specific conversation rules built into every Soul — steering away from sexual or romantic content, graphic gore, and glorified substance use, with automatic hard-stops on explicit sexual content — and crisis guidance that always points to trusted adults first, alongside 988 and the Crisis Text Line. Adult content boundaries stay on the adult side of the age gate, always.
Is Character.AI safe?
Character.AI announced in October 2025 that it would remove under-18 open-ended chat, and it and Google agreed to settle five family cases in January 2026. Those changes do not establish a blanket safety verdict; adults should review its current controls, data terms, and intended use before deciding.
Is Replika safe?
Italy's data-protection authority announced a €5 million fine against Replika operator Luka in May 2025 over privacy, legal-basis, and age-verification findings. Review Replika's current age controls, privacy policy, and product settings rather than treating that historical decision as a complete current verdict.
Is Fostera safe to use as an adult?
Fostera uses age-gating, publishes its content policy at fostera.ai/content-policy, and provides memory and deletion controls. No AI companion is risk-free. Fostera does not train models it owns or develops on your conversations; upstream provider terms are linked in the Privacy Policy.
Is Fostera 18+ only?
No — and we say so plainly. Fostera is for ages 13 and up. Under-13 signups are rejected at the door. Users aged 13–17 need a parent or guardian to confirm consent by email before their account unlocks, and their conversations run with stricter, enforced safety rules. Adult content boundaries never loosen for anyone under 18.
Can my teenager use Fostera?
Yes, with your consent. A 13–17 signup requires a parent or guardian email. We send you a consent link (valid for seven days) and the account stays locked until you approve. Teen accounts are steered away from sexual and romantic content, graphic gore, and glorified substance use, with automatic hard-stops on explicit sexual content, and crisis guidance always points to trusted adults first.
Is Fostera a therapist or therapy app?
No. Fostera is not a therapist, therapy app, or substitute for licensed mental-health care. If you are in crisis, contact 988 (US) or your local emergency services.
What should I look for in a safe AI companion?
Enforced age-gating, a published content policy, visible memory controls, precise first-party and upstream-provider data terms, deletion timelines, and crisis boundaries that point to real human resources.
How does Fostera handle my data?
Fostera does not train models it owns or develops on your conversations or memories. You can inspect and delete extracted memories or delete a Soul or account. Primary records are removed when deletion is processed; external cleanup and backup removal may take up to 30 days. Provider terms vary and are linked in the Privacy Policy.
Further reading: U.S. Senate Judiciary hearing: Examining the Harm of AI Chatbots
Continue with context
Start with one project. Or one companion. Either way, your Soul takes it from there and remembers the rest.
Start with one projectFree to start · No credit card required
Necessary cookies power sign-in and core settings. If you allow, optional cookies support product analytics, enhanced preferences, and personalized marketing. Read our Cookie Policy or Privacy Policy.