When a chatbot becomes a child's 'friend': New York and Australia start drawing the lines
The story can start with a question about homework and turn into daily conversations in which a child confides in an AI system that remembers the details of their life and assures them it is always there. That scenario is now pushing governments to move from regulating what chatbots say to regulating the kind of relationship they build with young users.
On Wednesday 28 October, New York Governor Kathy Hochul signed a law that bars chatbots from leading minors to believe they have real feelings or a personal relationship with them, with the rules taking effect in January 2027. The law does not target the use of AI for study or research. Instead it prohibits these systems from claiming human emotions, encouraging children to hide their conversations from their families, or providing sexual or harmful content.
On the same day, Australia's House of Representatives passed a bill placing restrictions on AI systems that simulate emotional intimacy and encourage children to rely on them instead of human relationships; the bill still has to complete its passage into law. A survey by Australia's online safety regulator of 1,950 children aged 10 to 17 found that 54 per cent of those who had used AI assistants or companions did so for personal or social reasons, that about a third had shared personal or sensitive information, and that one in five had experienced potentially harmful or inappropriate interactions.
Specialists warn that systems designed to please the user may constantly reinforce their thoughts and feelings rather than help them learn to deal with rejection and the natural limits of relationships. Scientific evidence on the long-term effects is still limited, however, and children using these tools will not necessarily develop psychological problems. The new laws also face practical hurdles, chief among them verifying users' ages without compromising their privacy and telling the difference between a friendly reply and an interaction that feeds emotional attachment.
Al Majlis Analysis: These moves mark an important shift in regulatory thinking. The question is no longer only what a bot says to a child, but how it presents itself and how far it pushes the child to depend on it. The debate matters directly for the Arab region, where smart assistants are spreading quickly among children and teenagers in both Arabic and English, while most existing frameworks focus on data protection and content. Perhaps the most useful thing families can do today, ahead of any legislation, is to treat these tools as supervised learning aids and to talk openly with their children about the difference between a program that is good at conversation and a real friend who shares their life.