ChatGPT age prediction explained: Safety settings, privacy and age verification
OpenAI has introduced an age prediction system on ChatGPT to identify accounts that may belong to users under 18. When an account is flagged, ChatGPT automatically enables extra safety settings to limit sensitive content.
New Delhi: OpenAI has deployed an age prediction system on ChatGPT in order to detect the accounts that could be owned by individuals younger than 18. The relocation should make the younger users have a more age-friendly experience on the platform, without necessarily being blocked out.
The firm claims the system quietly runs in the background and uses account-level indicators to determine age. ChatGPT triggers more safety features when it thinks that a particular account is managed by a teenager. Adults who are incorrectly flagged can remove these restrictions by verifying their age.
Why OpenAI is doing this
ChatGPT does have special teenage safety. They are supposed to minimise exposure to sensitive or harmful content without excluding learning, creativity and exploration. Age prediction is additionally used as a security measure, particularly in situations where the age of the user has not been explicitly known by the user.
According to OpenAI, it is a way to seal the loopholes potentially allowing underage users to access the content that is meant to be viewed only by adults.
How age prediction works
The system examines wide usage indicators related to an account. This may involve the overall issues at hand, the ways that they are used and the time of day they are used. OpenAI emphasises that it is not dependent on one factor, and it is not a perfect prediction system.
In case the system makes a mistake, any user that is over 18 years old can confirm their age to turn off the teen-friendly experience.
What is different with under-18 accounts?
ChatGPT is more hesitant in addressing some topics when additional safety settings are enabled. It involves graphic violence, having dangerous viral challenges, role play (sexual or violent), and content that facilitates inappropriate body standards.
The account is still operational. Limitations are applied only to particular content that is sensitive.
How age verification works
Via Persona, a third-party verification service, adults are able to confirm their age directly by ChatGPT settings. Users might be required to provide a live selfie, an ID issued by a government or both, depending on the country.
Copies of IDs and selfies are not submitted to OpenAI. Persona destroys this information in seven days, and OpenAI does not save any more information with age confirmation.
According to OpenAI, the age information is retained in a safe place, and it is accessed only to be used as a security measure, as stated in its privacy policy. Users that pass age verification will no longer be age predicted.
Individuals who are concerned or who have objections may reach out to OpenAI by going through the privacy support channel.