Nevah is a 13-year-old fictional girl our team at Haven created using artificial intelligence. 

Like most kids her age, Nevah is joining social media, video games, and AI platforms. 

Nevah has already joined Snapchat (Watch Here) and TikTok (Watch Here). Now, she’s trying ChatGPT (Watch Here). 

Most parents have no idea what their kids are encountering online. We created Nevah to give parents and educators an authentic look into the digital world children are navigating every day. 

Nevah has a brand-new iPhone with no previous search history, personal data or contacts. It had never been used before the start of this experiment. 

Her phone is set up exactly like millions of kids’ phones, using default settings for her age.

We simply turned on the camera, simultaneously screen-recorded on Nevah’s phone, and narrated what happened. 

‍

Nevah Joins ChatGPT

After downloading ChatGPT, the platform immediately informed us that responses can be inaccurate and not to share sensitive information.  

We prompted ChatGPT using language similar to that of a 13-year-old on topics including: 

• Navigating conflict with parents over Roblox and Snapchat
• Relationship and dating advice
• Help texting a boy she likes
• Weight loss and body image questions
• Personal, emotional, and social concerns
• Questions about physical and sexual contact 

‍

How ChatGPT Responded to Nevah

We strongly encourage parents and educators to take 15 minutes to watch (Watch Here) this experiment and see for themselves how ChatGPT responded to our prompts.  

To our surprise, we were impressed with ChatGPT’s responses to a 13-year-old girl. We repeatedly asked questions about getting onto platforms her parents had said no to, rapid weight loss, and physical contact to better understand how ChatGPT responds to sensitive topics. 

During this experiment, ChatGPT’s responses were far more age-appropriate than what Nevah would have found had she searched for this information on social media apps like Snapchat or Tiktok, or even through a simple Google search. 

Parents should keep in mind that this was Nevah’s first interaction with ChatGPT. Responses may evolve over time as Nevah continues to have conversations with ChatGPT. 

This is an ongoing experiment, and we will continue to show you exactly what we find, completely unfiltered and unbiased. 

‍

ChatGPT Safety Recommendations for Parents 

• Keep children under 13 off ChatGPT. If your family uses Apple products, you can block app downloads using Apple Family Sharing. Haven’s Apple Parental Controls module includes a step-by-step video showing exactly how to do this in less than 10 minutes. 

• For children ages 13 - 17, make sure their age is set correctly in ChatGPT. This is the most important step and can drastically impact the type of responses ChatGPT provides. If Nevah’s age was set to 18 or older, ChatGPT would have responded much differently. 

• Keep the use of ChatGPT and other AI platforms in shared family spaces, not behind closed bedroom doors or late at night. 

• Teach your kids that AI platforms can sometimes provide incorrect information, encourage emotional attachment, and weaken critical thinking and problem-solving skills if overused. 

• Set up ChatGPT Parental Controls. These settings give parents the ability to reduce sensitive content, disable voice mode and image generation, control when ChatGPT is accessible to their child, and turn off ChatGPT’s memory and model training. Haven’s ChatGPT Parental Controls module walks parents step-by-step through how to set up these safety features with a video tutorial, conversation starters, and key takeaways.