What parents should know about generative AI and child development

The rise of artificial intelligence (AI) has created passionate debate. It’s said to promise riches beyond imagining. It’s also said to threaten human existence itself.

Whatever the opinion, AI has already created a tectonic shift in society.

From financial markets, medicine, and education to news media and the fine arts, virtually every domain of human endeavour has already been impacted, if not transformed, by the rise of the machines.

This includes parenting. As AI tools and apps become increasingly used in family households, parents are faced with considering the impact on their children’s development.   

The risks and benefits of generative AI

Parents have already found many uses for AI. What’s the attraction? According to a recent U.S. survey, 90 percent of parents who use it believe it makes their job easier overall, and 91 percent say AI makes them feel more confident.

This shouldn’t be a surprise. From a parenting standpoint, AI tools can be a godsend, helping with daily family logistics, providing coaching in child development, and monitoring kids’ online safety. But AI is not without risks when it comes to our children.

A recent policy briefing paper, Mapping of GenAI impacts on child development [PDF], documents the emerging scientific consensus on the anticipated benefits and risks of generative AI (GenAI) in particular.

Released by the iRaise Coalition and authored by everyone.AI and Children and Screens, the brief was written as a response to the G7 call for science examining GenAI’s impacts on minors. It includes work from researchers across the globe who have looked at the opportunities and risks associated with GenAI.

Traditional AI vs. generative AI

To understand the risks and opportunities, it’s important to understand first the difference between GenAI and “traditional” AI.

Traditional AI applications, such as Siri, basically respond to user queries and interactions with pre-programmed answers and recommendations that human programmers have vetted. GenAI is less predictable. Through its “generative” capacity, it analyzes masses of information provided, including interactions with users, then creates responses to questions and prompts according to its own interpretations and biases.

This unpredictable quality is largely what creates concern for the interactions taking place between GenAI and children. And this is where parents need to be alert.

What the research says about kids and GenAI

The briefing paper examines the effects of GenAI across four developmental periods: 0-3, 3-6, 6-12, and 12-18 years. It looks at specific types of AI use at these ages to see what risks and benefits arise from children’s interactions with the technology.

The purpose is not to treat every AI exposure as harmful or beneficial, but to identify where GenAI may support child development and where it might disrupt it by becoming a substitute for capacities still forming in children. From there, the goal is to identify where expert guidance is needed to manage GenAI in a beneficial way for children.

AI and kids 0 to 3 years old

From birth to age three, the concern is how GenAI impacts the earliest interactions between infants and toddlers and their parents and caregivers.

At this early age, critical child development is taking place. Major elements include emotional bonding (attachment with parents and caregivers), early language, and sensorimotor learning. Through interactions with caring adults, children are learning trust, emotional regulation, communication, agency, and early mental models of how the world works.

AI tools may support this type of development indirectly. Baby monitors, sleep trackers, milestone apps, and parenting assistants supported by GenAI models may help caregivers respond better to children’s needs. However, when these systems draw attention away from the child or mediate caregiver-child exchanges, they risk interfering with the developmental processes they are intended to support.

Parents must ask whether use of AI during this period is disrupting their children’s development of trust, agency, language, and bonding with their own mothers and fathers.

AI and kids 3 to 6 years old  

Between ages three and six, children are developing their own abilities in symbolic play, language, emotion labelling, and recognizing the thoughts, feelings, and intentions of others. At the same time, they’re still developing the ability to distinguish what’s real from what’s fictional or imagined.

This is where significant danger exists. Children at this age easily accept non-human voices or characters as if they have minds and feelings. Consequently, they’re especially sensitive to human-like AI systems that are fluent, responsive, personalized, or emotionally expressive.

Parents need to consider how their children’s interactions with GenAI shape their understanding of the world at this age. While AI might support language and play in adult-guided contexts, it can also displace real-world interactions and blur the boundary between fiction and reality.

AI and kids 6 to 12 years old

Between six and 12 years, the early foundations that children have established in language, play, and social understanding are supposed to start developing into primary tools for their learning and socialization.

Research shows that children’s use of GenAI has become common at this age. Most of the 400,000 educational apps available today target this age group, and half of families in North America and China say their kids use learning apps multiple times a week. Children are using AI tools for their homework, creativity, games, and peer communication.

The question is whether AI strengthens or weakens the development of their own thinking, learning, and social abilities. 

The design features and default settings of GenAI apps are decisive. The same features that make AI tools useful can also reduce children’s cognitive effort, weaken their self-regulation, and displace their interactions with peers.

Questions remain about GenAI going forward

At every stage of child development, the concerns regarding GenAI are the same. The impacts depend on the capacities children are meant to be building at each age, and whether their interactions with AI systems are curtailing their development.

It seems AI can offer plausible benefits when it’s used to scaffold learning, creativity, and communication. But only if it preserves children’s own effort and social engagement.

The risk of harm arises when AI becomes a substitute for children’s own emerging capacities and interferes with their development—automating too much of their own cognitive or relational work, optimizing engagement with the AI instead of their peers, or becoming their primary companion, authority, or source of emotional reliance.

For policymakers tasked with protecting the public interest, the challenge is to create developmentally specific standards for AI and its use among children. The aim should be to create a clear distinction between supportive scaffolding and harmful interference, leading to practical safeguards and guidelines for AI product design and governance, but this is not guaranteed.

GenAI is a powerful and disruptive technology. How it impacts on our children, and ultimately our society, will depend on the safeguards we put in place. Understanding the capacities your child is meant to be building at each age and recognizing the risks and benefits can help parents to make their own choices and use AI wisely.