OpenAI builds a child safety team to protect young users

If you have kids or younger siblings, there is a good chance they are already talking to ChatGPT. And they are definitely not just using it to cheat on their history essays.

More and more teens are leaning on chatbots for mental health advice, social drama, and family issues. Now, OpenAI is officially building a defense line around them.

OpenAI just posted a job opening on its careers page, revealing the formation of a dedicated Child Safety team. The company is actively searching for a child safety enforcement specialist to study and prevent the abuse or misuse of its AI tools by underage users.

Policing sensitive content and youth incidents

According to the listing, this new group will not be working in a silo.

The Child Safety team will partner directly with OpenAI’s platform policy, legal, and investigation groups, as well as outside organizations, to manage processes, safety incidents, and content reviews tied to younger audiences.

The specialist will be tasked with enforcing OpenAI’s policies against AI-generated material and overseeing review systems for sensitive, kid-related content.

On paper, OpenAI already has rules: children under 13 are banned entirely, and teens aged 13 to 18 need parental consent. But as every tech platform knows, teenagers rarely read terms of service before clicking sign up.

The reality of kids turning to AI

The rush to put safeguards in place comes down to hard numbers.

A poll by the Center for Democracy and Technology found that 29% of kids have used ChatGPT to cope with anxiety or mental health struggles. Another 22% turned to it for issues with friends, and 16% used it to navigate family conflicts.

At the same time, harmful use is widespread. A survey by the UK Safer Internet Centre revealed that 53% of young people have seen peers use generative AI negatively, such as generating believable fake news or hurtful images to upset others.

Global watchdogs are taking notice. Audrey Azoulay, Director-General of UNESCO, recently urged governments to mandate age limits and data protections, warning that generative AI cannot enter education without proper safeguards.

Why it matters

OpenAI is learning the painful lesson every major consumer tech platform faces: eventually, you have to police your users.

For parents and teachers, this move is long overdue. Trying to ban ChatGPT in classrooms did not work, and kids are already forming emotional bonds with conversational bots. Having dedicated specialists inside OpenAI to handle edge cases and stop dangerous output is a necessary first step.

For OpenAI’s business, this is pure risk management. The company is aggressively pushing into schools, recently signing its first enterprise education customer and partnering with Common Sense Media. But running afoul of child protection rules like COPPA comes with crippling fines and reputational disaster.

That said, we think hiring human reviewers is just a band-aid. The real threat is not just bad prompts; it is the illusion of empathy in models talking to impressionable minds. Building real guardrails into the training data will be the true test.