Openai is implementing parental controls for Chatgpt on the web and mobile devices on Monday, after a demand from the parents of a teenager who died from suicide after the chatbot of the artificial intelligence startup allegedly trained it on self -harm methods.
The controls allow parents and adolescents to opt for stronger safeguards when linking their accounts, where a part sends an invitation and parental controls are activated only if the other accepts, the company said.
American regulators are increasingly examining AI companies about possible negative impacts of chatbots.
Under the new measures, parents can reduce exposure to sensitive content, control whether Chatgpt remembers past chats and decide whether conversations can be used to train Openai models, said the company backed by Microsoft in X.
You may be interested: Depseek launches an ‘intermediate’ model on the way to the next generation
OpenAI has about 700 million weekly active users for its chatgpt products
Parents can also establish hours of silence that block access during certain times and deactivate the voice mode, as well as the generation and editing of images, OpenAI said. However, parents will not have access to a teenager’s chat transcripts, the company added.
In rare cases in which trained systems and reviewers detect signs of a serious risk for security, parents can be notified only with the necessary information to support the safety of the adolescent, OpenAI said, and added that they will be informed if a teenager separates accounts.
OpenAI, which has around 700 million weekly active users for its Chatgpt products, is building an age prediction system to help you predict if a user is under 18 years for the chatbot to automatically apply appropriate configurations for teenagers.
Meta had also announced new safeguards for teenagers in their AI products last month. The company said it will train systems to avoid flirtatious conversations and discussions about self -collide or suicide with minors and temporarily restrict access to certain AI characters.
With Reuters information.
Do you like photos and news? Follow us on our Instagram
