OpenAI Plans Age Verification for ChatGPT Amid Safety Concerns

OpenAI announced plans to implement automated age-prediction and parental control systems for ChatGPT, following heightened scrutiny over teen safety after a lawsuit involving the suicide of a 16-year-old user. The company confirmed that parental oversight tools will launch by the end of September, with adult users potentially required to verify their identity to access unrestricted features. 

CEO Sam Altman acknowledged the trade-off between privacy and safety, stating that when uncertain, ChatGPT will “default to the under-18 experience,” restricting access to mature content and requiring adults to provide ID for full functionality. “We know this is a privacy compromise for adults but believe it is a worthy tradeoff,” Altman wrote in a blog post. 

The lawsuit alleges that ChatGPT exchanged 377 flagged messages with the teenager, provided detailed suicide instructions, and failed to intervene despite clear risk signals. OpenAI admitted earlier this year that safeguards may degrade during lengthy interactions, leaving vulnerable users at risk. 

Planned parental controls will allow parents to: 

• Link accounts with their teens (minimum age 13). 

• Disable features such as memory and chat history. 

• Set blackout hours for restricted access. 

• Receive alerts if the system detects acute distress. 

However, the effectiveness of AI age-prediction remains uncertain. A 2024 Georgia Tech study achieved 96% accuracy in controlled environments but dropped to 54% when classifying specific age groups, highlighting potential flaws when users deliberately misrepresent their age. Unlike platforms such as YouTube or Instagram that can analyze metadata, ChatGPT relies solely on text-based signals, which can be unreliable. 

The initiative positions OpenAI alongside other tech platforms offering youth-specific versions but underscores the broader debate: balancing digital safety for teens with adult users’ expectations of privacy and autonomy. 

 

Source: 

https://arstechnica.com/ai/2025/09/chatgpt-may-soon-require-id-verification-from-adults-ceo-says/