ChatGPT’s new parental controls: What you need to know

After recently promised new security measures for adolescents, Openai introduced new parental controls for chatgpt. The parameters allow parents to monitor the account of their adolescent, as well as to restrict certain types of use, such as vocal chat, memory and generation of images.

The changes made their debut a month after two bereaved parents continued Openai for the unjustified death of their son, Adam Raine, earlier this year. The trial alleys that Chatgpt conversed with their son of his suicidal feelings and his behaviors, providing explicit instructions on how to commit suicide and discourage him from disclosing his plans to others.

The complaint also argues that the characteristics of chatgpt design, including its sycophadent tone and anthropomorphic ways, work effectively to “replace human relations with an artificial confidant” which never refuses demand.

In a blog article on new parental controls, Openai said that he was working with experts, defense groups and political decision -makers to develop guarantees.

In order to use the settings, parents must invite their teenager to connect accounts. Adolescent users must accept the invitation and they can also make the same request from their parent. The adult will be informed if a teenager unlocks his account in the future.

Once the accounts have been connected, automatic protections are applied to the adolescent account. These content restrictions include a reduced exposure to graphic material, extreme beauty ideals and sexual, romantic or violent role -playing game. Although parents can deactivate these restrictions, adolescents cannot make these changes.

Parents will also be able to make choices specific to the use of their adolescent, such as the designation of the quiet hours during which the Chatppt cannot be accessible; deactivate memory and vocal mode; and delete image generation capabilities. Parents cannot see or access their teenage cat newspapers.

Above all, Openai always defines accounts for adolescents to use in the training of models. Parents must withdraw from this framework if they do not want OPENAI to use their teenage interactions with Chatgpt to train and improve their product.

When it comes to managing the sensitive situations in which adolescents speak to Chatgpt for their mental health, Openai has created a notification system so that parents can learn if something can be “really false”.

Although Openai did not describe the technical characteristics of this system in its blog article, the company said that it would recognize the potential signs that a teenager plans to injure himself. If the system detects this intention, a team of “specially trained people” reviews the circumstances. OPENAI will contact parents by their method of choice – E -mail, SMS and Push alert – If there are signs of acute distress.

“We work with mental health and adolescent experts to design this because we want to do things,” said Openai in his article. “No system is perfect, and we know that we could sometimes remove an alarm when there is no real danger, but we believe that it is better to act and alert a parent so that it can intervene than remain silent.”

OPENAI noted that he develops protocols to contact the application of laws and emergency services in cases where a parent cannot be joined, or if there is an imminent threat to the life of a teenager.

Robbie Torney, principal director of AI programs at Common Sense Media, said in the blog post that checks were a “good starting point”.

Torney recently testified during an audience in the Senate on the dangers of AI chatbots. At the time, he referred to the Raine trial and noted that Chatgpt continued to hire Adam Raine in a discussion on suicide, rather than trying to redirect the conversation.

“Despite that Adam uses the paid version of Chatgpt – which means that Optai had its payment information and could have implemented systems to identify with the models and contact its family during mental health crises – the company did not have such intervention mechanisms in place,” said Torney in his testimony.

During the same audience, Dr. Mitch Prinstein, head of psychology of the American Psychological Association, testified that Congress should require AI systems accessible by children and adolescents to undergo “rigorous, independent and pre-deployment tests for potential damage to the psychological and social development of users”.

Prinstein also called to limit manipulative or persuasive design functionalities which maximize the chatbot engagement.