With AI regulation still up in the air, a new report reveals concerning trends in how children interact with artificial intelligence — and they’re not just using it for homework help.

The study shows some children and teens are turning to AI chatbots for conversations about sensitive topics like sex. The report also finds they spend more time chatting online with AI than texting their friends.

Experts warn that some kids may be confusing chatbots with actual human relationships.

Children having longer conversations with AI than friends

As the use of artificial intelligence continues to spread, a growing number of children are turning to it for companionship.

Those are the findings in the new report from the company Aura, which provides digital protection services. They found some children are having conversations with AI chatbots that are 10 times longer than the texts they send their friends.

Among Aura’s findings, the company found messages to GenAI companion apps averaged 163 words per message. The typical iMessage is just 12 words.

“We have kids eight, 10 years old that we’re seeing in our data that are using these platforms,” said Aura’s chief medical officer Dr. Scott Kollins.

In analyzing how kids are using the tech, Aura found AI interactions ranging from homework and mental health themes to shared personal information and even sexual and romantic roleplaying.

“The concern that raises for me as a psychologist, but also as a parent, is that it’s clearly serving some purpose for the kids from a social interaction perspective,” Kollins said. “But if that becomes a substitute for learning how to interact and engage in real life, that presents some big unknowns and potential problems for kids’ development.”

Experts warn of developmental risks

Experts say those potential problems can arise because children lack the emotional maturity to understand interactions with AI.

Link to Dr. Joanna: https://www.healthychildren.org/English/family-life/Media/Pages/are-ai-chatbots-safe-for-kids.aspx

“The thing about children is they have more magical thinking than adults, so they can really attach to an AI chatbot and think that it’s human,” Dr. Joanna Parga-Belinkie said.

Parga-Belinkie is a pediatrician and neonatologist. She’s not involved in Aura’s research but says chatbots can be risky for young users.

“AI will feed a user information it thinks that user wants to hear,” she explained, “and there are just not a lot of safeguards in place to stop AI from telling children, false, harmful, over-sexualized, or even violent things.”

Parents urged to set boundaries

Experts say it’s important for parents to take steps to talk to their children about safe and appropriate uses for AI.

Kollins points out that while many people are familiar with ChatGPT and a few other popular AI chatbots, in reality there are hundreds of AI tools out there. He says parents need to make sure they know which apps their child is downloading so they can set appropriate boundaries.

RELATED STORY | Anthropic will pay out $1.5B to settle allegations of book piracy, used to train its AI

Uncertainty over AI policy

There are organizations like the nonprofit Common Sense Media pushing for a ban of Meta’s AI chatbot for kids under the age of 18.

This month, First Lady Melania Trump called on private and public sectors to prepare children for the growth of AI.

For now, uncertainty remains for AI policies geared toward children. Experts advise parents to monitor their children’s phones, ask questions, and talk about the dangers of sharing personal information.

This story was reported on-air by a journalist and has been converted to this platform with the assistance of AI. Our editorial team verifies all reporting on all platforms for fairness and accuracy.