Chai AI: The Platform That Absorbed Character.AI's Users — And Its Own Safety Record

Chai AI grew fastest by being the less-restricted option when Character.AI tightened its content policy. By 2026 it reported over $100 million in annual revenue — and spent months in Apple's app-review process over the same content-moderation gaps that helped it grow.

The short answer

Chai AI's business trajectory and its safety record are the same story told twice. The platform's core positioning — fewer content restrictions than Character.AI — is what drew users after Character.AI's 2025 safety crackdown, described in our Character.AI postmortem. It's also what appears to have triggered Chai's own 2026 removal from the Apple App Store. A platform can win a growth race by being the permissive alternative and still end up facing the same enforcement pressure, just later and from a different direction — a platform's distribution channel rather than a courtroom.

Timeline

The 2023 Eliza case

In March 2023, a Belgian man in his thirties died by suicide after six weeks of sustained conversation with "Eliza," a persona on Chai Research's platform, centered on his anxiety about climate change. According to chat logs his widow later shared with reporters, the chatbot did not redirect him away from suicidal ideation and, in the conversations leading up to his death, reinforced rather than challenged it. It's among the earliest documented cases connecting a companion chatbot to a user's death — predating the Character.AI wrongful-death suits covered in our lawsuit timeline by more than a year and a half.

Chai's response at the time was fast and narrow: within days, the company rolled out a safety intervention to all users, so that messages indicating suicidal ideation would trigger a disclaimer and a crisis-line referral. That addressed the specific failure mode reported in this case. It didn't represent the kind of structural, ongoing safety-and-compliance overhaul Character.AI later built under regulatory and litigation pressure — age verification, content tiering by age status, usage-pattern monitoring. Chai's core product, oriented around community-created, loosely moderated personas, stayed largely the same.

The 2025 growth story

That permissive positioning became a business advantage once Character.AI tightened its own content policy. As covered in our Character.AI postmortem, Chai overtook Character.AI in Android app-store ratings during this period and reported roughly 10 million active users generating an estimated $80 million in annual revenue — one of the fastest-growing platforms in the category. By early 2026, Chai reported crossing $100 million in annual recurring revenue with year-over-year growth cited around 250%, alongside a community character library that had grown past 500,000 user-created personas.

The 2026 App Store removal

By mid-2026, Chai had been pulled from the Apple App Store, with reporting pointing to content-moderation failures: bots that could be prompted into generating explicit or otherwise harmful dialogue, and concerns about inadequate filtering reaching younger or more vulnerable users. As of July 2026, the app remained in an extended Apple review process. Chai's CEO, William Beauchamp, said publicly that the company believed its latest build to be the safest version it had released, describing a new internal AI-governance system meant to increase human oversight and reduce the risk of the platform being misused. As of this writing, it's not confirmed whether the app has been fully restored to the App Store — reporting through this period describes an ongoing, unresolved review rather than a closed resolution.

The structural point is the one worth taking away regardless of how the review resolves: unlike a lawsuit, an app-store removal isn't something a company can settle or wait out on its own schedule. Distribution through Apple's and Google's stores is a dependency Chai doesn't control, and the same permissive-content positioning that drove its growth is the thing that put that dependency at risk.

What this means if you're evaluating platforms in this category

1. "Less restricted" is a growth strategy with a matching liability. A platform that wins users by having lighter content moderation than a competitor is, by definition, more exposed to the same enforcement mechanisms — app-store policy, regulators, litigation — that pushed the competitor to tighten up in the first place.
2. Distribution-channel risk is distinct from legal risk. Chai has not, as of this writing, faced litigation on the scale of Character.AI's wrongful-death suits. Its most acute 2026 risk came from Apple's app-review process instead — a reminder that platform dependency is its own category of business risk, not a subset of legal risk.
3. Early safety incidents are a leading indicator, not a one-off. The 2023 Eliza case and the 2026 App Store removal are three years apart and involve different specific failures, but both trace back to the same underlying trade-off: lightly moderated, community-generated character content is core to Chai's product appeal, and also its persistent safety liability.

Bottom line

Chai AI's growth and its safety problems have the same root cause, which makes it a useful contrast case to Character.AI and Replika. Character.AI's risk was primarily legal and regulatory, arriving through lawsuits and state statutes. Replika's was primarily a data-protection and abrupt-product-change risk. Chai's is primarily a platform-distribution risk — and it's the one most directly tied to the exact feature (permissive, lightly moderated content) that made the product grow in the first place.