OpenAI to predict user age from ChatGPT usage for verification

OpenAI is rolling out an age verification program for ChatGPT and says the service can often verify users from signals already linked to their accounts, such as the general topics they discuss or the times of day they use the tool. The company acknowledges the system is not perfect and can get it wrong.

The verification determines whether a user bypasses certain safety restrictions, including blocks on graphic violence or gore, viral challenges that could encourage risky behaviour, sexual, romantic, or violent role play, and content that promotes extreme beauty standards, unhealthy dieting, or body shaming.

In Italy, users have 60 days to complete verification when prompted, or certain features will be disabled. Even if you entered your birthday when creating an account, OpenAI may still seek verification. If it cannot infer age from past behaviour, users can verify through Persona, a third-party partner.

Italy

openai, chatgpt, age verification, account signals, usage patterns, safety restrictions, graphic violence, viral challenges, persona, italy