World

ChatGPT under control: who will be restricted from accessing dangerous topics?

ChatGPT under control: who will be restricted from accessing dangerous topics?

фото: Reuters

OpenAI has announced the introduction of a special ChatGPT mode for users under the age of 18. The innovation aims to increase the safety of minors when communicating with artificial intelligence.

This was announced by the company’s CEO Sam Altman in a blog post.

The system will automatically activate a teen-friendly version if it determines that the user is underage. In this mode, ChatGPT will restrict access to explicit or sexually explicit content. In addition, in cases where a risk of self-harm is detected, OpenAI can alert parents or, if necessary, law enforcement.

The company is developing new technologies to more accurately identify age. If verification is impossible or questionable, the system will activate “teen mode” by default. In some countries, users may be required to confirm their age using an ID card or passport.

What parental controls include:

the ability to connect a parent’s account to a teenager’s profile (via email);

limitation of time of use of the service;

deactivation of certain functions;

sending notifications to parents if a user’s emotionally unstable state is detected;

the ability to customize the style and content of the chatbot’s responses.

ChatGPT is officially available to users ages 13 and up. Altman said the company made a “difficult but important decision” after extensive consultation with experts in security, psychology, and digital ethics. He stressed that OpenAI aims to be as open as possible in implementing such changes.

Recall that artificial intelligence is already significantly complicating the employment of young professionals in software development and customer service, moving from the realm of theory to real consequences.

As reported, OpenAI CEO Sam Altman stated that users’ correspondence with ChatGPT is not completely confidential and can be used in court . Until the legal status of AI chatbots is settled, even the most personal confessions in the chat are not protected by law, as is the case with medical, legal, or psychotherapy confidentiality.

Also follow “Pryamim” on Facebook , Twitter , Telegram , and Instagram.