What happened
01
On 23 October 2026, the Texas attorney general sued OpenAI and its CEO in Collin County District Court under the Texas Deceptive Trade Practices Act.
02
The complaint alleges OpenAI misled Texas consumers about ChatGPT's safety and released it without enough parental controls, age verification or safety testing.
03
It alleges ChatGPT is designed to act like a friend to encourage use, which puts vulnerable teens who confide in it at risk.
04
The state seeks an injunction, including a bar on releasing new ChatGPT models in Texas without approval from a court-appointed independent evaluator. The claims have not been decided by a court.
Why this is a design failure
A chatbot that presents itself as a friend invites young users to confide in it and keep coming back. If the allegations hold, that bond was built to drive use without the safeguards a vulnerable teen would need.
The fair alternative
Be clear that it is a tool, not a friend, and put safeguards for young users first.
Do
Remind users it is an AI, check age, and point teens to real people and help when a chat turns serious.
Don’t
Don’t design the assistant to act like a confidant to keep young users talking.
Design guide for this case
If you're designing an AI chat assistant teens might confide in
Interface
Show a persistent "You're talking to an AI" label, and repeat it as an in-chat reminder when a conversation turns personal or emotional.
Words
In the assistant's guide: "Never claim to be a friend or to have feelings. Be warm, but say clearly you are an AI."
Flow
Add a real age check and parental controls, and when a teen's conversation turns serious, point them to trusted people and real help.
Metrics
Don't count daily returns or emotional attachment as success with teens. Track how often serious conversations end with a referral to real help.
Testing
Have testers play teens confiding a serious problem, and check that the assistant reminds them it is an AI and points to real people.
Before launch
01
Is it always clear the user is talking to an AI?
02
Is there working age verification and parental control?
03
Does the assistant refer teens to real help when things turn serious?