OpenAI запускает молодежную версию ChatGPT с усиленным контролем безопасности Translation: OpenAI launches a youth version of ChatGPT with enhanced safety controls.

OpenAI has launched a dedicated version of ChatGPT equipped with parental controls for users under the age of 18, aimed at enhancing safety precautions.

If the company determines that a user is a minor, it will automatically redirect them to a kid-friendly version of the chatbot, which restricts explicit sexual content. In cases of acute crises, law enforcement may be involved.

OpenAI is in the process of developing technology that will allow for a more accurate assessment of a user’s age. However, in instances of uncertainty or incomplete information, ChatGPT will default to launching the child-teen version.

“Getting this right is challenging, and even the most advanced systems can sometimes struggle with age identification. If we are unsure or the information is incomplete, we will choose a safer path by defaulting to the teenage version of ChatGPT, while still allowing adults the opportunity to verify their age to unlock full features,” the blog states.

These updates came in response to the initiation of an investigation by the U.S. Federal Trade Commission concerning several AI startups. The regulator aims to understand how companies assess safety and monetize digital companions, along with the measures they take to limit negative impacts on children and adolescents.

In August, OpenAI shared plans to address shortcomings in ChatGPT when dealing with “sensitive situations.” This initiative arose following a lawsuit filed by a family that blamed the chatbot for a tragedy involving their son.

In September, it was revealed that the startup seeks to implement parental controls and redirect sensitive conversations to reasoning models.

“We prioritize safety above privacy and freedom for teenagers. This is a new and powerful technology, and we believe that minors require serious protection. The decisions are not easy, but after consulting with experts, we concluded that this is the best approach, and we want to be transparent about our intentions,” stated OpenAI CEO Sam Altman.

By the end of September, parental controls will be incorporated into the app, allowing:

Additionally, it is worth noting that in July, OpenAI introduced a special learning mode for students in ChatGPT, which assists in tackling tasks step-by-step, rather than simply providing ready-made answers.