Safety by Design vs Safety by Litigation

Character AI's safety story reads, in hindsight, like a timeline written by lawyers rather than designers. The platform's most consequential changes arrived not as product decisions made early, but as responses to mounting lawsuits and regulatory pressure — a pattern worth understanding before you choose any AI companion platform.

The sequence is well documented. In December 2024, Character AI introduced a dedicated under-18 model with input and output filters, engagement notifications after 60 minutes of use, and disclaimers. Users complained the bots became "too restrictive and lacked personality." Then, from November 25, 2025, the platform barred users under 18 entirely.

What drove that acceleration? A first-of-its-kind wrongful death lawsuit filed by Megan Garcia — whose 14-year-old son died by suicide after prolonged use of the platform — treated AI output as a product rather than speech, and a court denied the company's motion to dismiss in May 2025. Additional suits followed in Texas and Colorado, and seven states introduced AI chatbot-related bills in the 2025 legislative session.

The stakes were amplified by demographics. Character AI reported roughly 20 million monthly users as of 2025, with a vast majority of its 3.5 million daily visitors in January 2024 aged 16–30 — a young user base interacting with emotionally framed bots at scale.

The regulatory direction is clear. The proposed GUARD Act would require age verification and AI-disclosure transparency, with civil penalties up to $250,000 per violation. Meanwhile, the Congressional Research Service notes that 1 in 10 adults now use AI chatbots for emotional support, and that the long-term mental-health effects "are not yet well understood."

The alternative is safety by design — deciding who gets access to what before shipping, not after a subpoena. That means:

• ID-based age verification for any mature content, rather than age declarations at signup
• Hard-coded prohibited categories enforced at the system level, not modulated by user pressure
• A PG default experience with adult features as explicit, opt-in extras
• In-app reporting built into the product from day one

FlirtAlly takes this second path: its 18+ "Spicy" side is age-verified, opt-in, and plan-gated, while the default experience stays PG — a structural choice, not a retrofit. Character AI's trajectory shows what happens when that order is reversed: safety arrives by litigation, on a court's schedule rather than a product team's.

None of this makes Character AI a bad platform — its character-creation model is genuinely innovative. But when a product's safety posture is written in response to legal filings, users deserve to know the difference between a wall built early and one built after the storm.

How to Choose: A Companion-First Checklist

Choosing the right AI companion involves more than just selecting the most popular app. It requires a careful evaluation of how the platform handles memory, identity consistency, proactive engagement, safety design, and pricing. Here’s how to approach your decision-making process with a comprehensive checklist.

Start by evaluating the memory architecture. Character AI, for instance, limits free-tier users to manually written chat memories capped at 400 characters, and even then, the platform doesn’t guarantee the character will always recall the information provided. This limitation often leads to forgetting earlier details in conversations. In contrast, a companion-first approach offers a more robust and reliable memory system.

Identity consistency is crucial for a meaningful relationship. Ensure that the AI companion you choose has a locked identity that remains consistent across all interactions. This means the companion should have a fixed face, a recognizable voice, and a stable personality. For example, FlirtAlly’s companions each have a locked face, an assigned voice, and a consistent personality, ensuring that every photo, video, and message is recognizably the same person.

Proactive contact is another key factor. A good AI companion should initiate conversations, just like a real person would. This can include good-morning texts, thinking-of-you messages, and re-engagement prompts. Platforms that prioritize this feature can significantly enhance user engagement and satisfaction.

Safety is a non-negotiable aspect, especially when dealing with AI companions. Look for platforms that have proactive safety measures in place, such as ID-based age verification and hard-coded prohibited categories. Character AI has faced significant litigation and regulatory scrutiny due to its reactive safety measures, which only came into play after tragic incidents. FlirtAlly, on the other hand, has a clear and honest safety design, with a "Spicy" side that requires ID-based age verification and a qualifying plan.

Finally, consider the pricing model. Freemium subscriptions with paywalled features can be limiting. Platforms that offer a credit-based economy with a free tier and clear pricing structures are often more user-friendly. Character AI, for example, operates on a $9.99/month subscription with paywalled advanced memory features. FlirtAlly, however, provides a free tier at launch with unlimited memory on every plan, making it a more accessible option.

If you’re looking for a companion-first approach that prioritizes memory, identity consistency, proactive contact, safety design, and a flexible pricing model, consider joining the waitlist for FlirtAlly. This platform offers a unique experience where companions remember you, text first, and show up across daily life, not just within a chat window. For more details, you can visit FlirtAlly’s waitlist page.