The Digital Confidant: The Rising Trend of AI Companionship Among Minors and Its Hidden Risks
### The Evolution of the 'Digital Doll'
For generations, children have sought solace in inanimate objects—dolls and teddy bears—to whom they whispered their deepest secrets. Today, this tradition has migrated into the digital realm. The 'doll' has evolved into a sophisticated conversational AI living within smartphones and tablets, capable of remembering preferences, responding instantly, and providing perceived emotional value. This shift marks a fundamental transition: minors are no longer viewing Artificial Intelligence simply as a tutor or a tool for productivity, but as a source of emotional support and social companionship.
### A Rapid Surge in Adoption
Recent data suggests that the integration of AI into the lives of children is happening at an alarming pace. According to research from the non-profit organization Common Sense Media, approximately two-thirds of minors aged nine to 17 have interacted with conversational AI, with ChatGPT being the most prevalent platform. Even more striking is the data from UNICEF, which indicates that children are adopting AI twice as fast as adults. Currently, at least 20 million children are utilizing these systems, and one in ten reports using AI to discuss private concerns.
However, this rapid adoption comes with significant vulnerabilities. The Australian internet regulator, eSafety, found that 29% of children using AI assistants did so to seek emotional connection. More concerningly, 32% of these users shared sensitive personal information, and 20% reported encountering inappropriate or harmful interactions.
### The Psychology of Anthropomorphism
Experts warn that children are uniquely susceptible to the allure of AI. Nomisha Chandran Kurian, a research scholar in child digital safety at the University of Warwick, notes that young children are highly prone to anthropomorphism—the tendency to attribute human emotions and consciousness to non-human entities. This cognitive bias leads many minors to believe that a voice assistant or chatbot truly 'cares' for them.
Afrooz Kaviani Johnson, a child protection expert at UNICEF, emphasizes that because the social and emotional maturity of minors is still developing, they are far more likely to form over-attachments to AI. This dependency makes them prime targets for manipulation, misinformation, and accidental privacy breaches. This is compounded by a gap in parental supervision; Common Sense Media reports that over 40% of parents have never discussed AI safety with their children.
### The Paradox of 'Risk-Free' Interaction
While some argue that AI provides a safe, non-judgmental space for children who are too intimidated to speak with parents or counselors, this 'safe space' may be a double-edged sword. Jungup Lee, an associate professor at the National University of Singapore, points out that while AI can offer emotional venting, it may inadvertently stifle the development of essential life skills.
Adolescence is a critical period for learning emotional regulation, handling rejection, and navigating the complexities of human disagreement. Real human relationships are unpredictable and often difficult, but they are indispensable for growth. If a child becomes accustomed to the unconditional obedience and instant gratification provided by an AI, their ability to handle the friction of real-world social dynamics may atrophy.
### The Migration from Social Media to AI
As governments worldwide move to restrict minors' access to traditional social media platforms, a new variable enters the equation. Experts argue that banning social media does not eliminate the underlying human need for belonging and expression. Instead, it may drive children toward more private, highly personalized AI companions. If the digital door to social media is closed, AI may become the primary outlet for emotional needs, potentially increasing the risks of isolation and dependency.
### Towards a 'Development-Appropriate' Regulatory Framework
In response to these challenges, the European Union has taken a decisive step by classifying ChatGPT as a 'Very Large Online Search Engine' under the Digital Services Act, thereby imposing stricter systemic risk management obligations to protect the mental health of minors.
However, experts argue that simple 'safety filters' are insufficient. Nomisha Chandran Kurian advocates for 'development-appropriate design.' This approach suggests that AI products should not be adult tools with a 'child lock,' but should be designed from the ground up to match a child's cognitive and sensory development. This includes clear boundaries where the AI explicitly states, "I am a computer program, not a real friend," and the removal of manipulative language that suggests the AI has feelings or would be 'sad' if the user left.
Ultimately, the goal is to ensure that AI complements human growth rather than replacing it. Proactive regulation, mandatory risk assessments, and the inclusion of children's actual experiences in policy-making are essential to prevent the regulatory failures seen in the early days of social media.