Neural networks are becoming a common tool for children, making it important to discuss safe usage rules within the family. Ahead of the new school year, experts from Kaspersky Lab have outlined the main risks associated with AI and provided several recommendations on how to keep your child safe.
A chatbot is not a best friend. A child may perceive a conversation with a chatbot as a private chat and consider AI a trustworthy interlocutor with whom they can share details about themselves, their family, school, or personal feelings. However, queries and dialogues may be stored by the service and, depending on its settings and privacy policy, used to improve models. That is why it is essential to explain to children that they should not share their name, home address, phone number, documents, or other personal information with smart chatbots. The same rules apply to other people's data: do not upload photos, messages, or documents without the owner's permission.
AI in studies: an assistant, not a doer. With the start of the school year, children may actively use neural networks for learning: asking to explain a complex topic, generate essay ideas, check text, solve homework, or prepare presentations. However, when a child can get a ready-made essay or solution in seconds with AI, there is a temptation to pass off the generated result as their own work without delving into the essence of the answer. Parents should agree with their children on rules: the neural network can help explain a topic, suggest a plan, provide examples, or ask practice questions—but not do the entire assignment.
'AI hallucinations.' One of the most common mistakes is treating a generated response as if it were a search engine result or an expert opinion. In reality, a language model generates a response based on patterns identified in the data it was trained on and the context provided. A convincingly worded answer is not necessarily accurate: the neural network may mix up facts, invent a source, cite a non-existent study, or confidently describe an event that never happened. Such misinformation is often called an 'AI hallucination.' Children should remember the rule: the more important the question, the more carefully the answer needs to be verified.
'AI can become a valuable tool for a child's learning if used to explore a topic, check understanding of study material, or refine a draft essay—but not to replace their own work. The key is to learn not to take every AI-generated answer at face value. At first glance, it may sound plausible but contain fabricated facts, false references, or erroneous reasoning,' explains Andrey Sidenko, Head of Child Online Safety at Kaspersky. 'Parents can help by encouraging children to cross-check important information with textbooks and reliable sources, explain answers in their own words, and ask for help if something seems unclear. Such habits will help children benefit from AI while developing critical thinking skills that will serve them not only during the school year but throughout their lives.'
'Teaching children the rules of proper AI usage is a key skill for living in a safe digital environment,' noted Mushvig Mammadov, official representative of Kaspersky in Azerbaijan. 'For example, at the start of the school year, parents can practice with their children: analyze the neural network's response to a homework assignment, check sources, discuss which parts of the text are useful, where there are inaccuracies, and what should be improved. Specialized security solutions will provide additional protection against phishing, scams, and other cyber threats.'
In addition to rules for safe and responsible AI use, it is important to follow basic digital security practices. To protect yourself and your children online, Kaspersky experts recommend:
• Enable two-factor authentication wherever possible;
