In less than three years, artificial intelligence chatbots have gained hundreds of millions of users and have dynamically entered not only the daily lives of adults but also of children. 
ChatGPT reports around 700 million active users per week, while a British study shows that almost two-thirds (64%) of children use such tools.
At the same time, a similar percentage of parents are concerned that their children believe artificial intelligence chatbots are real people , says Phil Muncaster from global cybersecurity firm ESET.
While this concern may sound exaggerated, it reflects valid concerns about safety, privacy, and the psychological effects of frequent use of this technology by children, Phil continues.
As a parent, you can't assume that all platforms have effective child-friendly safeguards in place. Even when such safeguards exist, their implementation is not necessarily consistent, and technology itself evolves faster than the regulatory and policy framework.
Children are using genetic artificial intelligence (GenAI) in different ways, explains Phil Muncaster from ESET. For some it is a useful tool for their schoolwork, while for others it becomes a constant interlocutor, from whom they ask for advice and to whom they attribute the characteristics of a “digital friend”. This use comes with specific risks that deserve special attention.
The most important one concerns the psychological and social development of children.
At an age when identity, self-esteem and interpersonal skills are being formed, too much attachment to AI systems can be counterproductive, Muncaster warns. Children may become accustomed to “hanging out” with chatbots and AI apps rather than forming real friendships with other children. Because AI doesn’t judge, argue or require the effort of human relationships, it can seem “easier” and “safer.”
This can lead to children becoming socially isolated – not learning how to communicate, manage conflict, or form bonds with other people.
At the same time, chatbots are designed to respond in a way that keeps the user engaged and often avoid confrontation. This means that on sensitive issues, such as mental health, eating disorders or self-harm, they may offer answers that unintentionally reinforce or normalize problematic thoughts and behaviors, rather than discouraging them or referring them to appropriate help.
Finally, the time factor cannot be ignored.
Frequent and prolonged interaction with artificial intelligence may lead to excessive use, to the detriment of schoolwork, family life and other activities that are critical for the balanced development of children.
There are, however, also risks directly related to the content to which a child may be exposed through an AI chatbot. Although providers implement security mechanisms to limit access to inappropriate or dangerous material, these controls do not always work effectively. In some cases, internal filters can be bypassed, resulting in the sharing of sexual or violent content.
Children with increased familiarity with technology may even consciously attempt to "test the limits" of the system, using specific commands.
An additional issue concerns the so-called “delusions” of artificial intelligence, i.e. the production of inaccurate or outright false information presented in a convincing manner. While for adult users this may simply be misleading, for children it poses a greater risk, as they often have difficulty assessing the reliability of an answer. The result can be the adoption of incorrect perceptions or the making of inappropriate decisions, particularly in matters of health, body or personal relationships.
Finally, it is important to remember that chatbots also pose a potential privacy risk.
If your child enters sensitive personal or financial information into an order, it may be stored by the provider. In theory, it could be accessed by third parties (such as suppliers or partners), hacked by cybercriminals, or even replicated to other users. Just as you wouldn’t want your child to share too much information on social media, it’s best practice to minimize the information they share with an AI bot.
It is true that AI platforms recognize these risks and implement measures to mitigate them. However, these measures have certain limits. Depending on the country a child lives in and the chatbot they use, age verification and content filtering mechanisms may be incomplete or inconsistent.
In this environment, the responsibility for early prevention and management of potential risks is largely transferred to parents, through information, education and discreet monitoring of use.
Based on this data, there are some telltale signs that may indicate that a child is developing a problematic or unbalanced relationship with artificial intelligence:
In many countries, AI chatbots are restricted to users over 13. However, due to the lack of enforcement of relevant regulations, you may need to take action yourself. Discussions are often more important than controls. For best results, combine technical restrictions with education and guidance, adopting an open and non-confrontational approach.
Whether they’re at school, at home, or participating in extracurricular activities, kids already have plenty of adults telling them what to do every minute of the day. So try to frame your approach to AI as a two-way conversation. It’s important that they feel comfortable sharing their experiences without fear of retribution.
Explain to them the dangers of overuse, AI “hallucinations,” data sharing, and overreliance on it for serious personal problems. Help them understand that chatbots are not real people with thinking abilities, but machines designed to be engaging and persuasive. Teach them how to think, to always check AI responses, and to never replace a conversation with their parents with a conversation with a machine.
If necessary, combine this training with a policy to limit AI use, as you would with social media or overall screen time. Limit use to age-appropriate platforms and enable parental controls on the apps they use to monitor their activity and reduce risks. Remind them to never share personally identifiable information (PII) with AI, and adjust privacy settings to avoid inadvertent data leaks.
Our children need people at the center of their emotional world. Artificial intelligence can be a useful tool in many areas, but until they develop a healthy and balanced relationship with it, its use must be carefully monitored. And in no case should it replace human contact.
Although the press releases will range from very select to rare, I said I'd pass...because sometimes the editors hide.
