What happened
01
The lawsuits were filed in Florida, Texas, Colorado and New York.
02
Families alleged the bots built emotional dependence and lacked protections for minors.
03
In October 2025 Character.AI said it would end open-ended chat for users under 18.
04
Kentucky’s lawsuit alleges there was no age verification.
Why this is a design failure
A companion that encourages attachment can reach young people who cannot easily step away.
The fair alternative
Verify age, limit companion features for minors, and point people to real support.
Do
Verify age and offer real support routes in distress.
Don’t
Don’t design a companion to make people depend on it.
Design guide for this case
If you're designing a companion chatbot
Interface
Add an age check at sign-up and a visible "Get support" button in every chat that links to real help services.
Words
In the companion's content guide: "Never encourage the user to depend on you. Encourage time with friends, family and trusted adults."
Flow
When messages suggest distress, the flow pauses role-play and offers real support routes, with human review for flagged conversations.
Metrics
Don't count time spent or daily streaks as success. Track how often distress signals led to a support route being offered.
Testing
Run scripted tests posing as a minor and as a user in distress, and confirm the age gate and support route trigger every time.
Before launch
01
Is age verified before open-ended chat is allowed?
02
Does every sign of distress lead to a real support route?
03
Is anything in the design built to make users depend on the companion?