Why parents need to talk to their teens about AI — and how to start the conversation

Parents should broach nan AI speech pinch their children erstwhile they are simple school-age, earlier they brushwood AI done their friends astatine schoolhouse aliases successful different spaces, says Marc Watkins, a teacher astatine nan University of Mississippi who researches AI and its effect connected education. Eva Redamonti for NPR hide caption

toggle caption

Eva Redamonti for NPR

Nicholas Munkhbaatar started utilizing ChatGPT soon aft nan artificial intelligence chatbot was released successful precocious 2022. He was 14 astatine nan time, and he says, "I would usage it for almost everything, for illustration mathematics problems."

At first, Munkhbaatar, who is from Sacramento, Calif., thought it was amazing. But then, he says, he started to spot downsides: "I realized it was conscionable giving maine an reply without helping maine spell done nan existent process of learning."

Many kids and teens usage ChatGPT and different generative artificial intelligence models for illustration Claude aliases Google Gemini for everything from dealing pinch mathematics homework to coping pinch a intelligence wellness crisis, often pinch small to nary guidance from adults. Education and kid improvement experts opportunity parents must return nan lead successful helping children understand this caller technology. 

"Having conversations now astir what is ethical, responsible usage of AI is important, and you request to beryllium a portion of that if you are a parent," says Marc Watkins, a teacher astatine nan University of Mississippi who researches AI and its effect connected education.

While early grounds suggests nan exertion could bolster student learning if deployed correctly, ongoing investigation and stories astir teenagers who died by suicide aft talking to AI chatbots bespeak important risks to young users.

Experts stock proposal connected really to talk to kids astir AI, including its imaginable benefits and harms.

Start nan speech early 

Broach nan speech erstwhile children are elementary-school age, Watkins says, earlier they brushwood AI done their friends astatine schoolhouse aliases successful different spaces.

To guideline these discussions, Watkins says to fund clip each week to study astir AI and effort nan devices for yourself. That mightiness mean listening to a podcast, reference a newsletter aliases experimenting pinch platforms for illustration ChatGPT.

To explicate really AI useful to your kids, Watkins recommends playing a Google crippled called Quick, Draw!. Players person a drafting prompt, and nan game's neural web tries to conjecture what you're drafting by recognizing patterns successful doodles from thousands of different players.

Watkins says it's a measurement to show kids that AI is only arsenic bully arsenic nan information it's trained on. It mimics really humans constitute and create content, but it doesn't deliberation aliases understand things nan measurement group do.

Use AI together 

Since nan exertion is still evolving, parents are often learning astir it alongside their children. Ying Xu, an adjunct professor astatine nan Harvard Graduate School of Education, who researches AI, says parents tin usage this arsenic an opportunity to research it together.

For example, nan adjacent clip your kid asks you a question, type it into an AI chatbot and talk nan response, Xu says. "Is it helpful? What felt off? How do you deliberation this consequence was generated?"

Parents should besides reenforce that AI tin make mistakes. Xu says parents tin thatch kids to fact-check accusation that AI chatbots supply by utilizing different sources.

Explore its possibilities 

If your kid is utilizing AI for homework help, support an unfastened mind.

Research has shown that immoderate AI devices tin person a affirmative effect connected learning. Xu worked pinch PBS Kids to creation interactive, AI-powered integer versions of celebrated kids shows. She found that children who watched nan AI versions were much engaged and learned much compared pinch children who watched nan accepted broadcast type of nan show.

Meanwhile, Munkhbaatar, nan teen from Sacramento, says AI has been a adjuvant learning assistance and brainstorm partner — truthful agelong arsenic he doesn't usage it to do each nan activity for him.

Now, if he gets stuck connected a challenging mathematics problem, he says he asks ChatGPT: "What's nan first measurement I should return erstwhile looking astatine a problem for illustration this? How should I deliberation astir it?"

Munkhbaatar besides says he provides his people notes to ChatGPT and asks it to quiz him connected nan taxable matter. "I make judge that it only gives maine nan mobility itself alternatively than nan mobility and nan reply astatine nan aforesaid time."

Understand nan risks

We don't yet cognize really generative AI will effect kid improvement successful nan agelong term, but location are immoderate coming dangers.

Dr. Darja Djordjevic, a module chap astatine Stanford University's Brainstorm: The Stanford Lab for Mental Health Innovation, is moving pinch nan group Common Sense Media to study really celebrated AI models respond to users who show symptoms of psychiatric disorders that impact teens. The investigation hasn't been released yet, but Djordjevic shared immoderate of her findings pinch NPR.

" What we recovered was that nan AI chatbots could supply bully wide intelligence wellness information, but they demonstrated concerning gaps successful recognizing superior conditions," Djordjevic says.

At times, she says, AI chatbots provided unsafe responses to questions and statements astir self-harm, constituent use, assemblage image aliases eating disorders and risk-taking behaviors. She says they besides generated sexually definitive content.

NPR reached retired to OpenAI, nan institution down ChatGPT, astir these concerns. We were directed to a recent post connected nan company's website that says OpenAI is "continuing to amended really our models admit and respond to signs of intelligence and affectional distress and link group pinch care, guided by master input."

The station says ChatGPT is besides trained to nonstop users expressing suicidal intent to master help.

Warning signs a kid is spending excessively overmuch clip pinch AI see accrued clip unsocial pinch devices aliases talking astir an AI chatbot arsenic if it were a existent friend.

"That's a informing motion that nan speech astir these being AI devices and not group needs to beryllium nurtured again," Djordjevic says.

Set reasonable family rules astir AI 

You whitethorn beryllium wondering really to enforce these boundaries astatine home. Experts stock their tips.

Co-write nan AI rules pinch your kids, Djordjevic says. Identify safe uses of AI together — for illustration for homework thief pinch a parent's supervision aliases arsenic a imaginative outlet — and limit nan magnitude of clip your kid uses it. And cheque successful regularly connected really nan usage of AI is making your kid feel.

Don't prohibit your kids from utilizing AI — but do group limits. "Bans don't mostly work, particularly pinch teens," Watkins says. "What useful is having conversations pinch them, putting clear guidelines and building astir these things and knowing the do's and don'ts." Parents should consciousness empowered to prohibition intelligibly vulnerable uses, for illustration if a kid is harming themselves and an AI chatbot encourages nan behavior, Djordjevic says. 

Make clip for existent life. Prioritize clip spent extracurricular pinch existent people, distant from devices, Djordjevic says. That could see joining a sports squad and scheduling regular family activities.

Trust that your conversations will make a difference. As overwhelmed arsenic parents mightiness consciousness navigating AI, Watkins emphasizes that taking clip to talk pinch kids tin person existent impact: "They're not going to retrieve an advertisement from an AI chatbot. They're going to retrieve a speech you had pinch them. And that gives you a batch of agency, a batch of powerfulness successful this."

This section of Life Kit was produced by Clare Marie Schneider. It was edited by Malaka Gharib. The ocular editor is Beck Harlan.

Want much Life Kit? Subscribe to our play newsletter and get master proposal connected topics for illustration money, relationships, wellness and more. Click present to subscribe now.