OpenAI is tightening control over teenagers' use of the ChatGPT service by adding a new mechanism. Under this mechanism, if a teenager's account is suspended for violating the company's policies on violent threats or committing violent acts, their parents can be informed of this.
This measure is part of the expansion of parental control tools in ChatGPT and aims to provide families with greater ability to intervene if they observe signs that could put the teenager or others at risk.
Notifying parents
Under the new system, parents who have linked their account to their teenager's account may receive a notification from OpenAI if the teenager's account is suspended by OpenAI due to violent activity.
OpenAI has explained that this notice can include an alert about an important update related to the teenager's account. Parents can choose the relevant option to view more details; including whether the account was deactivated for violating the company's policies on violent behavior.
However, OpenAI emphasized that these alerts are designed to be limited and specific and are not aimed at conversations related to fiction writing, gaming, news, politics, general expressions of anger, or theoretical questions.
Expansion of parental control tools
OpenAI had previously launched parental control tools in ChatGPT. These tools allow parents to link their account to their teenager's account and manage a set of settings related to how ChatGPT is used.
Among the features in this section are setting hours when ChatGPT is not available to the teenager, restricting sensitive content, disabling voice or memory capability, preventing image generation, and preventing the teenager's conversations from being used to train the model.
Parents can also enable Study Mode directly through the parental control settings. This feature, instead of providing a final answer, helps the teenager's learning process by offering tips and guidance.
Protecting others
The importance of this measure has grown amid increasing concerns about how AI systems handle violent content.
OpenAI had stated that after controversies arose about one of the suspects in the Tumbler Ridge mass shooting in Canada using ChatGPT, it strengthened its safety protocols. The company had also faced criticism for how it managed the account prior to the attack.
According to the company, the purpose of launching the new system is to give families the opportunity to intervene in a timely manner if they observe dangerous signs, rather than only facing the account being closed.
Part of a broader protective system
OpenAI's measures are not limited to addressing violent content. The company's parental control system also includes alerts about signs that may indicate a teenager is at risk of harming themselves.
When systems detect signs of acute distress, contacting parents is also possible. OpenAI emphasizes that these tools are not a substitute for parental supervision and intervention, but are designed to provide additional tools to understand how teenagers use AI and to help intervene when necessary.