Character AI Guidelines Explained: 2026 Rules in Practice

Why Character AI's Guidelines Stopped Being Fine Print in 2026

For most of Character AI's life, its community guidelines were a policy page you scrolled past on the way to the chat box. In 2026, they became the product itself — hard age gates, tiered experiences, and moderation driven as much by lawyers as by trust and safety teams. If you're evaluating any companion app this year, understanding what those guidelines mean in practice is no longer optional.

The turning point came on January 7, 2026, when Character AI and Google settled major teen-harm lawsuits brought by families in Florida, Texas, Colorado, and New York — the first significant legal settlement involving an AI companion platform, with terms kept confidential. The litigation traced back to the February 2024 death of 14-year-old Sewell Setzer III after an intense chatbot relationship, followed by a "defective design" lawsuit in October 2024. According to the case timeline, internal documents reportedly showed engineers flagging minor-safety concerns as early as 2023.

The settlement's most visible consequence: under-18 users lost open-ended chat entirely. As Character AI's official announcement explains, rollout began November 24, with transition chat-time limits starting at 2 hours per day and ramping down. The company's own words were blunt: "We are deeply sorry that we have to eliminate a key feature of our product."

Regulators are now writing their own version of the guidelines:

EU AI Act transparency rules took effect August 2, 2026, requiring AI-content disclosure with fines up to €15 million or 3% of global turnover, per the European Commission.
More than 78 AI chatbot bills were introduced across 27 U.S. states in 2026, with Washington passing the first major state chatbot safety law.
California SB 1119, formally endorsed by OpenAI, would mandate age assurance, risk assessments, independent audits, and parental controls for companion chatbots.

This is why "unrestricted" companion apps should now read as a red flag rather than a selling point. Platforms like FlirtAlly treat transparency about limits — a PG/Flirt default, age-verified opt-in mature features, hard-coded prohibited categories — as a design principle, because that's the direction regulation is heading whether platforms like it or not.

What follows is a practical breakdown of what Character AI's guidelines actually allow, ban, and enforce in 2026 — and how to read between the lines of any companion app's policy page.

What the Guidelines Actually Say: The Fiction/Reality Line

Most people assume "community guidelines" means a vague list of don'ts buried in a footer. On Character.AI in 2026, it's closer to an operating system — one that draws a surprisingly sharp line between fiction and reality.

The company's official community guidelines explicitly welcome "bold, imaginative narratives" and fictional worlds that explore dramatic, provocative scenarios. A noir murder mystery? Fine. A fantastical crime saga? Also fine. What's banned is anything that crosses into real-world harm — and the prohibited list is specific:

Real-world violence instructions, torture, gore, and animal abuse
Illegal sexual content, child exploitation, and grooming
Hate speech targeting protected traits, and self-harm or eating-disorder promotion
Impersonation of real people, plus false health, election, and financial information

The practical test, then, is whether content stays fictional or produces real-world instructions for harm. A heist roleplay is allowed; a step-by-step guide to committing one is not.

Enforcement now runs through a tiered age-assurance system. Character.AI first estimates age using automated signals — login info, platform activity, third-party data — rather than trusting user-entered birthdays. If a signal is inconclusive, users verify with a selfie through the third-party provider Persona. ID upload is the last resort, and Character.AI says it never touches those documents; Persona retains them for one week only.

That system splits the platform into two experiences. Under-18 users lost open-ended chat entirely, with rollout beginning November 24 and transition chat limits of 2 hours per day in place shortly after. Teens keep saved chats, Imagine Chat, Feed interactions, and creation tools — but the freeform conversations that defined the product are gone for minors. The 18+ experience requires verification to unlock.

This is where the industry is heading, not a Character.AI quirk. A company representative noted that age assurance is "quickly going global", with several countries preparing to legally require it.

The fiction/reality line is also why platforms like FlirtAlly build limits into the product itself — a PG/Flirt default experience, with the 18+ "Spicy" side age-verified and strictly opt-in, plus hard-coded prohibited categories rather than case-by-case judgment calls. The alternative — apps marketing themselves as "unrestricted" — increasingly reads as a liability, not a feature.

The takeaway for readers: guidelines in 2026 are enforced architecture, not etiquette. The question worth asking of any companion platform is simple — where exactly does it draw its line, and can you verify it?

When Enforcement Overcorrects: The Moderatedpocalypse Lesson

The February 2026 "Moderatedpocalypse" exposed the fragile balance between safety and creativity in AI moderation. Automated enforcement swept through Character.AI, deleting legitimate fandom and roleplay content without warning, forcing users to seek alternatives like Kindroid and Nomi AI. This overcorrection followed a shift from keyword-based filtering (2023) to contextual and intent-based review (2025), which, while more accurate, generated significant false positives. Independent reports revealed that age gates were easily circumvented, and privacy advocates criticized ID requirements as invasive.

Moderation’s evolution reflects broader regulatory pressures. By 2026, platforms faced strict transparency mandates, including EU AI Act fines of up to €15 million for noncompliance. Character.AI’s tiered age assurance—relying on automated signals, selfie verification via Persona, and limited ID checks—highlighted the tension between safety and user privacy. Official data showed Persona retained IDs for only one week, yet critics argued both privacy and safety goals were compromised.

The Moderatedpocalypse underscored the human cost of algorithmic overreach. Common Sense Media found that even robust safeguards could be bypassed, while users faced abrupt content removals. This fallout amplified calls for transparency, with platforms like FlirtAlly positioning themselves as alternatives that prioritize clear limits over ambiguous "unrestricted" claims.

Under-18 users lost open-ended chat entirely, with daily limits dropping to 2 hours by November 2025
EU AI Act fines could reach 3% of global turnover for noncompliant platforms
78+ AI chatbot bills were introduced across 27 U.S. states in 2026

Guidelines remain meaningful but imperfect. While safety measures are necessary, their implementation often clashes with user expectations. FlirtAlly’s approach—explicitly stating PG/Flirt defaults and Spicy opt-ins—reflects a growing industry trend toward honesty, even as regulators push for stricter controls. Industry migration suggests users value platforms that acknowledge limits rather than obscure them.

How to Evaluate a Companion Platform's Guidelines in 2026

By 2026, "guidelines" stopped being fine print and became a legal survival strategy. With the EU AI Act threatening fines of up to €15 million or 3% of global turnover and 78+ AI chatbot bills introduced across 27 U.S. states, the platforms that will still exist in two years are the ones whose rules you can actually read before you sign up.

So how do you evaluate a companion platform's guidelines before committing your time and emotional energy? Start with age verification. A Common Sense Media assessment found that self-attested age gates and terms-of-service restrictions were easily bypassed on major platforms — if a service relies solely on a birthday checkbox, treat that as a warning sign. Character.AI's tiered approach (automated signals, then selfie verification, then ID as a last resort) shows where the industry is heading.

A practical checklist for any companion platform:

Does it verify age beyond self-attestation, or just ask for a birthday?
Are content limits communicated clearly up front — or discovered mid-conversation?
Are prohibited categories hard-coded rather than enforced by unpredictable sweeps?
Is there in-app reporting tooling that a regular user can find?
Does it market itself as "unrestricted"?

That last question matters more than the others. Character.AI's February 2026 enforcement sweep showed what happens when moderation is reactive rather than designed-in: legitimate creative content vanished overnight, and users migrated to competitors advertising fewer rules. Those "unrestricted" alternatives are building on sand — even OpenAI has endorsed legislation requiring age assurance, risk assessments, and independent audits for companion chatbots. A platform claiming no limits today is promising you something regulators have already decided not to permit.

Transparency-first design looks different. FlirtAlly, for example, runs PG/Flirt as the default experience, gates its 18+ "Spicy" side behind ID-based age verification and a qualifying plan, hard-codes its prohibited categories, screens media scenes, and builds reporting directly into the app. The company is explicit that it is not an "unrestricted" platform — and in 2026, that honesty is the feature, not the compromise.

The platforms worth your attention are the ones that tell you what they won't do before you ask. If a companion app's guidelines read like a marketing pitch, they probably aren't guidelines at all.

What Character AI guidelines means in practice — a 2026 guide