OpenAI is introducing new ChatGPT safety features for users under 18, including teen accounts, voice controls, and reminders designed to encourage healthy AI use.
Teen users will be able to disable ChatGPT’s human-like voice responses, reducing the technology’s ability to sound and feel like a real person.
OpenAI will also provide regular reminders encouraging young people to take breaks, remember that ChatGPT is an AI system, and understand that it is acceptable to pause before continuing a conversation.
Although OpenAI said the changes were not introduced in response to specific reports of children believing ChatGPT is alive, concerns are growing that users may view interactive AI tools as conscious or sentient.
Mustafa Suleiman, Microsoft’s head of AI, said last year that AI products appearing to be sentient kept him “up at night” because of their potential impact on society.
In several other ways, OpenAI’s teen account system resembles protections already introduced by major social media platforms.
In its blog post announcing ChatGPT for Teens, external, OpenAI said teenage AI use should include protections that reflect young people’s development, strengthen real-world relationships, and promote healthy long-term use.
Teen accounts will be automatically assigned to users who identify themselves as teenagers during sign-up, as well as users OpenAI determines may be under 18.
However, the company has not explained how it will determine a user’s age or identify teen accounts.
ChatGPT is not intended for children under the age of 13.
Source: www.bbc.co.uk