A child needs more than an answer machine
A patient chatbot may explain a fact. It cannot become the whole environment in which a child learns to trust and question.
A child asks why the moon changes shape, and an assistant replies instantly. The explanation might be useful. Yet a learning moment can contain more than an answer: shared attention, a drawing, a second question, and an adult who notices what the child misunderstood.
A child's question may also be an invitation to spend time together. The answer matters, but so do the follow-up, the drawing, and the adult noticing what was not understood.
Children may attribute understanding, friendship, or authority to a system that responds conversationally. An engaging voice does not establish that the system knows the child, understands their situation, or is safe to consult about every subject. Age-appropriate design and adult involvement matter.
Use assistance for bounded activities: comparing an explanation with a book, generating ideas for a supervised experiment, or asking for a simpler description that an adult checks. Avoid entering identifying details or private concerns into tools without understanding how information is handled.
Useful assistance leaves the child more able to ask and check. It should not require private disclosure, replace every human listener, or make a smooth response seem beyond questioning.
Leave substantial room for physical play, human conversation, mistakes, and not knowing. These are not obsolete forms of learning. A child who can ask where an answer came from and notice when it does not make sense is gaining more than a child who merely collects smooth responses.
The ambition should be a richer learning environment, not a cheaper substitute for every attentive person in it. An answer can satisfy a question. A relationship can help a child become the kind of person who keeps asking.