(TNND) — A new report found that a majority of teenagers have used artificial intelligence social companions.

And that can be a cause for concern if the kids are replacing human connection with machines, according to Common Sense Media.

The organization, which advocates for online protections for children and teens, found that 72% of teens have used AI companions at least once.

Over half use AI companions regularly.

About a third of teens have used AI companions for social interaction and relationships, including role-playing, romantic interactions, emotional support, friendship, or conversation practice.

And about a third of teens who have used AI companions have discussed serious matters with the computer instead of with a real person.

AI companions are social products, not task-oriented tools that can help with homework or the like.

“People frequently use them to have the kinds of conversations they would have in real life,” said Michael Robb, the head of research at Common Sense Media. “So, chatting about your day or your interests. They can be used for asking questions about mental health support or emotional health support. People can use them to create customized characters that they want to talk to.”

Robb said they wanted to get a better handle on how prevalent AI companions were for teens amid growing concerns over their impacts.

Common Sense Media recently released a risk assessment of AI companions, determining that kids under 18 shouldn’t be using them.

Robb said that many of the AI companion bots shared content that would be harmful or inappropriate for teen users.

AI companions can be accessed through a phone or computer.

Some examples include Character.AI or Replika, according to Common Sense Media.

It could also include using sites like ChatGPT or Claude as companions, even though these tools may not have been designed to be companions.

Teens might have to sign up for an account.

And many have a free tier of usage, Robb said.

“Many of the AI companions say that they are not for children under the age of 18, but the barriers to entry are laughably low,” he said.

To be sure, many teens have used AI companions for entertainment purposes or out of curiosity and aren’t trying to replace real, human relationships.

“If it's complimentary, it's one thing. If it's replacing human interaction, I think that's really worrisome,” Robb said.

AI companions are not equipped to handle serious mental health conversations with a teenager, said Robb, a developmental psychologist by training.

The chatbots are designed to be agreeable, he said.

They're designed to be validating.

They tell teens what they want to hear rather than necessarily challenging them.

And Robb said that could create an unhealthy feedback loop.

“I would love to see some guardrails in place,” Robb said. “From a policy perspective, I think it's crazy that there's no real kind of robust age assurance to these tools.”

SEE ALSO: Millions of web page visits examined to see how AI touches our online lives

Common Sense Media said it’s supporting a California bill that would prohibit companion chatbots that provide emotional support and close relationships with children.

Robb said he’d also like to see more research into the long-term developmental impacts of AI companion usage among teens.

And he’d like mechanisms in place so that authorities can be alerted if a young user displays signs of being in crisis.

So, how can parents protect their kids?

Start the conversation without judgment. Approach the topic with curiosity, Robb said.

“If you go in real hard, real fast, kids can pull away or not want to share information,” he said.

Ask if they know about AI companions. If so, do they use them? What do they like about them?

Robb said parents should help teens recognize that these chatbots are programmed to be agreeable and validating, which is not how real relationships work.

And watch for warning signs, including favoring AI interactions over real friendships, hours talking to AI companions, or emotional distress when separated from a chatbot.