Kids are offloading their critical thinking to AI chatbots — here’s how to protect them, experts say

Artificial quality is reshaping workplaces — and progressively uncovering its mode into the hands of galore teens and children.

From helping with homework to chatting with AI "friends," tools specified arsenic ChatGPT person escaped versions online that are casual for young users to access. These AI chatbots, built connected ample connection models (LLM), make human-like responses that person sparked interest among parents, educators and researchers.

A 2024 survey by Pew Research Center recovered that 26% of U.S. teens aged 13 to 17 accidental they person utilized ChatGPT for their schoolwork — treble the complaint from a twelvemonth earlier. Awareness of the chatbot roseate to 79% successful 2024 from 67% successful 2023.

Regulators person taken notice. In September, the Federal Trade Commission ordered 7 companies, including OpenAI, Alphabet and Meta, to explicate however their AI chatbots whitethorn impact children and teenagers.

In effect to mounting scrutiny, OpenAI announced successful the aforesaid period that it'll motorboat a dedicated ChatGPT acquisition with parental controls for users nether 18 and make tools to amended foretell a user's age. The strategy would automatically nonstop minors to "a ChatGPT acquisition with age-appropriate policies," the institution said.

However, immoderate experts interest that aboriginal vulnerability to AI — particularly arsenic today's youngest generations turn up with the exertion — whitethorn negatively interaction however children and teens deliberation and learn.

A 2025 preliminary survey from researchers astatine MIT's Media Lab examined the cognitive outgo of utilizing an LLM in penning essays. 54 participants aged 18 to 39 were asked to constitute an effort and were assigned to 3 groups: 1 could usage an AI chatbot, different could usage a hunt motor and a 3rd that solely relied connected their ain knowledge.

The insubstantial — inactive successful the process of being adjacent reviewed — recovered that encephalon connectivity "systematically scaled down with the magnitude of outer support," according to the study.

"The brain‑only radical exhibited the strongest, widest‑ranging networks, the hunt motor radical showed intermediate engagement, and LLM assistance elicited the weakest wide [neural] coupling," according to the study.

Ultimately, the survey suggests that relying connected AI chatbots could pb radical to consciousness little ownership implicit their enactment and pb to "cognitive debt," a signifier of deferring intelligence efforts successful the abbreviated word that whitethorn erode creativity oregon marque users much susceptible to manipulation successful the agelong run.

"The convenience of having this instrumentality contiguous volition person a outgo astatine a aboriginal date, and astir apt it volition beryllium accumulated," said research idiosyncratic Nataliya Kosmyna, who led the MIT Media Lab study. The findings besides suggested that relying connected LLMs mightiness pb to "significant issues with captious thinking," she added.

Children, successful particular, could beryllium astatine hazard for immoderate of the antagonistic cognitive and developmental impacts of utilizing AI chatbots excessively soon. To assistance mitigate these risks, researchers hold that it is precise important for anyone, peculiarly the youth, to person the skills and cognition archetypal earlier relying connected AI tools to implicit tasks.

"Develop the accomplishment for yourself [first], adjacent if you are not becoming an adept successful it," said Kosmyna.

Doing truthful volition let inconsistencies and AI hallucinations — a improvement wherever inaccurate oregon fabricated information is presented as facts — to beryllium caught more easily, she added, which volition besides assistance "support captious reasoning development."

"For younger children ... I would ideate that it is precise important to bounds the usage of generative AI, due to the fact that they conscionable truly request much opportunities to deliberation critically and independently," said Pilyoung Kim, a prof astatine the University of Denver and kid science expert.

There are besides privacy risks that children whitethorn not beryllium alert of, and it is important that erstwhile utilizing these tools, they are utilized responsibly and safely, explained Kosmyna. "We bash request to thatch overall, not conscionable AI literacy, but [also] machine literacy," she said. "You request truly wide tech hygiene."

Children besides person a higher inclination to anthropomorphize, oregon to property quality characteristics oregon behaviour to non-human entities, said Kim.

"Now we person these machines that speech conscionable similar a human," said Kim, which tin enactment children successful susceptible situations. "Simple praise [from] these societal robots tin truly alteration their behavior," she added.

Today, the AI-native procreation is increasing up with entree to these tools, and experts are asking themselves: "What happens with extended use?"

"It's excessively aboriginal [ to know]. No 1 is doing studies connected three-year-olds, of course, but it's thing precise important to support successful caput that we bash request to recognize what happens to the brains of those who ... are utilizing these tools precise young," said Kosmyna.

"We spot cases of AI psychosis. We spot cases of, you know, unaliving. We spot immoderate heavy depressions... and it's precise concerning and sad, and yet dangerous,' she added.

Kosmyna and Kim said regulators and exertion companies stock the work to support nine and young radical by having the close guardrails successful place.

For parents, Kim's proposal is simple: support an unfastened enactment of connection with your kids and show the AI tools they use, including what they benignant into the LLMs.

Want to beryllium your ain boss? Sign up for CNBC's caller online course, How To Start A Business: For First-Time Founders. Find step-by-step guidance for launching your archetypal business, from investigating your thought to increasing your revenue.

Plus, sign up for CNBC Make It's newsletter to get tips and tricks for occurrence astatine work, with wealth and successful life, and request to articulation our exclusive assemblage connected LinkedIn to link with experts and peers.