Much of the public discourse around artificial intelligence has focused, understandably, on its potential to fundamentally alter the workforce. But we must pay equal attention to AI’s threat to fundamentally alter humanity — particularly as it continues to creep, unregulated, into early childhood.
AI may feel like a developing force largely disconnected from the way we raise children. The truth is, AI is already impacting children’s developing brains in profound ways. “Alexa” now appears in babies’ first vocabularies. Toddlers increasingly expect everyday objects to respond to voice commands — and grow frustrated when they don’t. And now, one of the world’s largest toy companies has launched a “strategic” partnership with OpenAI. Research shows that children as young as three can form social bonds with artificial conversational agents that closely resemble the ones they develop with real people.
The pace of industry innovation far outstrips the speed of research and regulation. And our kids’ wellbeing is not at the center of these inventions. Consider Meta’s chatbots, capable of engaging in sexually explicit exchanges — including while posing as minors — which are available to users of all ages. Or Google’s plans to launch an AI chatbot for children under 13, paired with a toothless disclaimer: “Your child may encounter content you don’t want them to see.”
Now, with the Senate negotiating a budget bill that would outright ban states from regulating AI for the next decade, parents stand to be left alone to navigate yet another grand social experiment conducted on their children — this time with graver circumstances than we’ve yet encountered.
As a pediatric physician and researcher who studies the science of brain development, I’ve watched with alarm as the pace of AI deployment outstrips our understanding of its effects. Nowhere is that more risky than in early childhood, when the brain is most vulnerable to outside influence. We simply do not yet know the impact of introducing young brains to responsive AI. The most likely outcome is that it offers genuine benefits alongside unforeseen risks; risks as severe as the fundamental distortion of children’s cognitive development.
This double-edged sword may sound familiar to anyone versed in the damage that social media has wrought on a generation of young people. Research has consistently identified troubling patterns in adolescent brain development associated with extensive technology use, such as changes in attention networks, reward processing pathways similar to behavioral dependencies, and impaired face-to-face social skill development.
Social media offered the illusion of connection, but left many adolescents lonelier and more anxious. Chatbot “friends” may follow the same arc — only this time, the cost isn’t just emotional detachment, but a failure to build the capacity for real connection in the first place.
What’s at stake for young children is even more profound. Infants and young children aren’t just learning to navigate human connection like teenagers, they’re building their very capacity for it. The difference is crucial: Teenagers’ social development was altered by technology; young children’s social development could be hijacked by it.
To be clear, I view some of AI’s potential with optimism and hope, frankly, for the relief they might provide to new, overburdened parents. As a pediatric surgeon specializing in cochlear implantation, I believe deeply in the power of technology to bolster the human experience.
The wearable smart monitor that tracks an infant’s every breath and movement might allow a new mom with postpartum anxiety to finally get the sleep she desperately needs. The social robot that is programmed to converse with a toddler might mean that child receives two, five or ten times the language interaction he could ever hope to receive from his loving but overextended caretakers. And that exposure might fuel the creation of billions of new neural connections in his developing brain, just as serve-and-return exchanges with adults are known to.
But here’s the thing: It might not. It might not help wire the brain at all. Or, even worse, it might wire developing brains away from connecting at all to another human.
We might not even notice what’s being displaced at first. I have no trouble believing that some of these tools, with their perfect language models and ideally timed engagements, will, in fact, help children learn and grow — perhaps even faster than before. But with each interaction delegated to AI, with each moment of messy human connection replaced by algorithmic efficiency, we’re unknowingly altering the very foundations of how children learn to be human.
This is what keeps me up at night. My research has helped me understand just how profoundly important attachment is to the developing brain. In fact, the infant brain has evolved over millennia to learn from the imperfect, emotionally rich dance of human interaction: the microsecond delays in response, the complex layering of emotional and verbal communication that occurs in even the simplest parent-child exchange. These inefficiencies aren’t bugs in childhood development, they’re the features that build empathy and resilience.
It is safe to say the stakes are high. Navigating this next period of history will require parents to exercise thoughtful discernment. Rather than making a single, binary choice about AI’s role in their lives and homes, parents will navigate hundreds of smaller decisions. My advice for parents is this: Consider those technologies that bolster adult-child interactions. Refuse, at least for the time being, those that replace you. A smart crib that analyzes sleep patterns and suggests the optimal bedtime, leading to happier evenings with more books and snuggles? Consider it! An interactive teddy bear that does the bedtime reading for you? Maybe not.
But parents need more than advice. Parents need, and deserve, coordinated action. That means robust, well-funded research into AI’s effects on developing brains. It means regulation that puts child safety ahead of market speed. It means age restrictions, transparency in data use, and independent testing before these tools ever reach a nursery or classroom.
Every time we replace a human with AI, we risk rewiring how a child relates to the world. And the youngest minds — those still building the scaffolding for empathy, trust and connection—are the most vulnerable of all. The choices we make now will determine whether AI becomes a transformative gift to human development, or its most profound threat.
Dana Suskind, MD, is the founder and co-director of the TMW Center for Early Learning + Public Health; founding director of the Pediatric Cochlear Implant Program; and professor of surgery and pediatrics at the University of Chicago.