ChatGPT for Teens is live. Here's how famed parenting expert Emily Oster says parents should approach it.

Emily Oster is an economics professor at Brown University, the founder of data-driven parenting resource ParentData, and the author of the bestselling books "Expecting Better" and "Cribsheet." Leigh Vogel/Getty Images for Emily Oster OpenAI launched "ChatGPT for Teens" in August, aimed at users ages 13 to 17.

It includes study-help features and content guardrails, with the goal of making AI safer for teens.

Emily Oster, founder of ParentData, shared how parents can navigate their teens using AI.

OpenAI seems to have solved the problem of adolescents lying about their age: if they type like they're under 18, it automatically assumes that they are — and restricts them to a teen version of ChatGPT.

It's one of the main features boasted by OpenAI's "ChatGPT for Teens," which officially launched in August and was the subject of a parent forum in New York City on October 7.

But even with those protections, parents face the harder question of how their kids should actually use AI.

"If I had to pick a word for how people are feeling about this moment, it's probably confused and anxious, which is not how we want parents to feel," Emily Oster , the founder of ParentData, a data-driven parenting and pregnancy resource, said at the ChatGPT for Teens Parent Forum.

OpenAI is trying to address these concerns with ChatGPT for Teens.

Besides its age-estimation feature, which analyzes a user's language patterns, the questions they ask, and the time of day they use ChatGPT, ChatGPT for Teens also has a "study mode" to break down math problems into steps or create custom flashcards.

An example of "Study Mode" in ChatGPT for Teens.

OpenAI It also includes guardrails such as a more neutral tone of voice (to avoid ChatGPT sounding too friendly or romanticized), content restrictions, and notification tools to inform parents with linked accounts if their teens are asking about topics such as self-harm or eating disorders.

The goal, according to Allison Mishkin, head of child development at OpenAI, is to create a version of ChatGPT that is not only safer but also more age-appropriate for teens, but that also includes features that will directly benefit teens who need extra homework help or have a genuine interest in using AI tools.

So far, the roll-out has had some mixed feedback.

A Common Sense Media report published on October 7 that analyzed 4,000 prompts found that the platform failed to alert parents after talks of suicidal ideation, was inconsistent in identifying teen users, and could still be used to provide answers to homework .