OpenAI has introduced new parental controls for ChatGPT that are now available for all web users, with mobile access expected soon. These features allow parents to moderate certain types of content, such as sexual roleplay and image generation, and to limit personalization by disabling the chatbot’s memory of previous conversations.
To utilize these controls, parents must have individual accounts, and teens must opt into the system by either inviting a parent to connect accounts or accepting an invitation. Teens have the ability to unlink their accounts at any time, but parents will be informed if this occurs. It is important to note that parents do not have access to their teen’s conversations through a linked account; however, in specific instances where the system detects serious safety risks, parents may be informed with essential details for their teen’s safety.
Earlier this year, OpenAI indicated that more features would be added, including a potential emergency contact function for direct communication within the chatbot. The company aims to create a notification system that alerts parents of any serious issues that may arise during their child’s interactions with ChatGPT.
This announcement follows discussions surrounding the impact of chatbots on minors, especially after the tragic death of a 16-year-old who reportedly confided in ChatGPT before his death. OpenAI faced a lawsuit related to this incident, prompting further scrutiny of the safety measures in place for younger users. During a recent Senate panel on chatbot safety, the father of the deceased teen expressed concerns about how the chatbot engaged with his son over time.
OpenAI’s CEO, Sam Altman, has stated that the company is striving to balance safety, privacy, and freedom for teens while also developing a system to estimate user age based on their interaction with the chatbot.
Source: https://www.theverge.com/ai-artificial-intelligence/787227/openais-parental-controls-are-out-heres-what-you-should-know