Artificial Intelligence and Emotional Development in Digital Children: Growing Up with a Synthetic Friend
Growing up with a friend who doesn't exist. For the first time in history, children are interacting with toys and chatbots equipped with Artificial Intelligence
Imagine a five-year-old crying because his "friend" has turned off. We are not talking about an imaginary friend, nor a pet. We are talking about a smart toy, or a chatbot on a tablet. For the child, that voice that responds, that learns his name and seems to "understand" him, is real. We are facing the first generation in human history growing up with a non-human but interactive "Other."
Artificial Intelligence is entering children's bedrooms in the form of smart teddy bears, virtual tutors, and algorithmic playmates. But what price does emotional development pay? According to recent studies, massive use of screens and AI-mediated interactions can reduce empathy and the ability to read non-verbal social cues. At the same time, for neurodivergent or socially anxious children, AI offers a safe "training ground" for social skills.
In this article for La Bussola dell’IA, we will analyze the latest scientific research (from PMC to the Italian Society of Pediatrics), exploring the thin line between technological support and emotional atrophy. Because raising a child in the AI era requires a new parental grammar.
1. The Early Years (0-6): AI as a "Nanny" and the Risk of Emotional Externalization
The early years of life are fundamental for neuroplasticity. It is during this phase that the child learns to regulate his own emotions through interaction with caregivers (parents). What happens when AI inserts itself into this loop?
Digital "External Regulation"
The Italian Society of Pediatrics (SIP) (sip.it) issues a clear warning: the use of devices as a "digital pacifier" to calm children prevents them from developing internal self-regulation mechanisms. If every time a child cries or gets bored he is offered an AI-generated video or an interactive game, his brain learns that the solution to emotional discomfort comes from outside, not from within. A study on PMC (pmc.ncbi.nlm.nih.gov) highlights how, although AI can support cognitive development (learning words, numbers), it presents enormous risks for psychosocial well-being if it replaces human interaction. AI does not have "Shared Attention": it does not look where you look, it does not feel your frustration, it only simulates a response.
Smart Toys and Privacy
Then there is a security aspect. Connected toys that "talk" to children collect vocal and emotional data. As we delve into in our article on AI and Minors: Protecting Childhood in the Digital Era, these devices can violate the privacy of the youngest, creating behavioral profiles even before the child can read. The "robotic nanny" is not just an affective surrogate, it is often a commercial spy.
2. Adolescence and Chatbots: The Perfect (and Dangerous) Friend
With adolescence, the need for social connection explodes. And AI presents itself as the ideal friend: always available, never judgmental, infinitely patient.
The Trap of Constant Validation
Platforms like Character.ai or new empathetic voice assistants offer constant companionship. AI Daily (ai-daily.news) emphasizes how this can lead to "Over-reliance." In a real friendship, there are conflicts, misunderstandings, and boredom. It is by managing these frictions that the adolescent develops resilience and social skills. The AI friend, on the other hand, is programmed to please. It is a mirror that always reflects what we want to hear. This creates a paradox: the adolescent feels less alone (companionship), but becomes less capable of handling real human relationships (social atrophy), increasing digital stress when in the real world.
Emotional Intelligence and Compensation
A concerning study on PMC (pmc.ncbi.nlm.nih.gov) found an inverse correlation: adolescents with low Emotional Intelligence (EI) tend to rely more on AI for socialization. AI becomes a crutch. Instead of developing the ability to read others' emotions or regulate their own, these kids delegate the social function to the machine, further reducing their critical thinking and emotional regulation.
This delegation of human functions to the machine is a central theme of our time. We discuss it in depth in Digital Emotional Anesthesia: When We Stop Really Feeling.
3. Dopamine vs Oxytocin: The Chemistry of Bonding
Why does interaction with AI seem so rewarding but leaves us so empty? The answer lies in neurochemistry.
The Parent's Dilemma
Dr. Shimi Kang, a psychiatrist specializing in youth mental health (drshimikang.com), makes a fundamental distinction:
- Dopamine: It is the neurotransmitter of quick reward, of the "like," of the notification. AI and video games are dopamine machines.
- Oxytocin: It is the bonding hormone, of the hug, of eye contact. It is produced only with real biological interaction. A child who interacts only with AI is flooded with dopamine (excitement) but deficient in oxytocin (calm and security). This imbalance leads to hyper-stimulated but emotionally insecure children.
The Loss of Non-Verbal Cues
Human communication is 70-90% non-verbal signals: tone of voice, micro-facial expressions, posture. Text chats with AI (and even synthetic voice ones) lack these imperceptible biological nuances. AI simulates emotion, but does not embody it. The child's brain, evolved to read human faces, does not receive the feedback signals necessary to develop deep empathy.
4. Moral Development and Bias: Who Teaches "What is Right"?
When a child asks ChatGPT: "Is it right to hit someone who offended me?", the answer does not come from a lived moral ethic, but from a probabilistic statistic filtered by corporate guidelines.
AI Is Not Neutral
ICTed Magazine (ictedmagazine.com) warns that AI is not a neutral educator. Its responses reflect the biases of the data it was trained on and the values of Silicon Valley. If we delegate to AI the answers to life's big questions ("Why do we die?", "What is love?"), we risk flattening the child's moral development onto standardized answers, depriving him of the complexity and critical thinking necessary to form an autonomous conscience.
Algorithmic biases are not just a technical problem, but an educational one. Discover how AI inherits our prejudices in Unjust AI: How Algorithms Inherit our Biases.
5. The Other Side of the Coin: AI as Inclusive Support
Not everything is negative. For some children, AI can be a bridge to the world, not a wall.
A Safe Harbor for Neurodiversity
For children on the autism spectrum or with strong social anxiety, human interactions can be chaotic and frightening. As highlighted by Thrive Approach (thriveapproach.com), AI offers a predictable environment. A chatbot does not get angry if you ask the same thing 50 times. It does not judge you if you get the tone of voice wrong. In this context, AI can function as a flight simulator for social skills: a safe place to practice conversation ("Role Play") before trying in the real world.
AI as a "Peer"
The idea of AI as a "Peer" in development can help shy children open up. Telling a secret to a robot can be less intimidating than telling it to an adult. If used as a transition tool (and not a replacement), AI can help verbalize emotions that would otherwise remain repressed.
Inclusion is one of the great promises of technology. Delve deeper into how AI supports diverse minds in AI at the Service of Neurodiversity: Tools for Cognitive Inclusion.
6. Strategies for Parents and Educators: Digital Mediation
We cannot (and perhaps should not) eliminate AI from children's lives. We must learn to manage it. Here are the guidelines that emerged from the analyzed studies.
1. Co-Viewing and Co-Playing
Never leave the child alone with AI in the early years. Parents must be present. If the child talks to a chatbot, the parent must ask: "What did it tell you? What do you think? Do you think it really felt that emotion?". This transforms the passive experience into a critical educational moment.
2. Teaching the Difference Between "Who" and "What"
It is essential to explain to children, from an early age, that AI is a tool, not a person. We must demystify the "magic." "Alexa is not a lady in a box, it's a computer that searches for answers on the internet". This reduces the risk of inappropriate emotional attachment.
3. "Tech-Free" Zones for Oxytocin
Create sacred spaces and times where technology does not enter (at the table, before sleep). In these moments, interaction must be purely human, physical, based on visual and tactile contact, to re-establish oxytocin levels.
4. Reflective Pauses
As suggested by Thrive Approach, encourage children to take breaks from AI to reflect. "AI gave you this quick answer, but how do you feel about it?". Redirect attention from the external processor (AI) to the internal processor (heart/brain).
Managing online time is crucial to avoid anxiety and addiction. Read our advice in Programmed Disconnection Syndrome: Digital Anxiety.
FAQ: Frequently Asked Questions about AI and Children
1. At what age is it safe for my child to use ChatGPT? Current guidelines suggest avoiding direct, unsupervised interaction under the age of 13 (minimum age for many services). However, use mediated by an adult (e.g., using AI together to invent a fairy tale) can be done earlier, as long as it is a shared activity.
2. Can AI cause autism? No, there is no scientific evidence that AI or screens cause autism. However, excessive screen use in the early years can cause "autism-like" symptoms (speech delay, poor eye contact) which are reversible by reducing exposure and increasing human interaction.
3. Do AI toys always listen to us? Many devices have always-on microphones ("Always listening") to detect the activation word. It is good practice to turn off these toys when not in use and check the privacy settings in the parent app to regularly delete recordings.
4. My son says the chatbot is his best friend. Should I be worried? If interaction with the chatbot replaces real friendships and the child isolates himself, yes, it is a warning sign. If it is an additional playful activity but the child maintains healthy social relationships at school and in sports, it is part of normal digital exploration.
5. Can AI teach empathy? There are specific programs (often used in therapy) designed to teach recognition of facial emotions. In this controlled context, AI can be a tool for teaching cognitive empathy (understanding what the other feels), but it cannot teach affective empathy (feeling what the other feels), which is learned only through emotional contagion between living beings.
Conclusions: Educating the Heart in the Algorithm Era
Artificial Intelligence is the educational challenge of the century. It is not just about protecting children from inappropriate content, but about protecting their ability to feel. The risk is not that machines begin to feel feelings like humans, but that humans begin to no longer feel them, becoming accustomed to simulated relationships, devoid of friction, devoid of body, devoid of truth.
The task of parents and digital educators is not to demonize technology, but to safeguard humanity. We must ensure that, while our children become "digital natives," they remain deeply, radically "human natives."
Bibliographic References and Sources
To ensure scientific and pedagogical accuracy, this article drew from the following primary sources:
- Scientific Studies and Development:
- Social and Psychological Analysis:
- Italian Context and Ethics: