Children are beginning to use artificial intelligence alongside the messaging, video, social and gaming platforms that already shape their digital lives.
That creates a new challenge for parents, schools and cybersecurity providers. Teaching children not to click suspicious links remains important, but it no longer covers the full range of decisions they make online.
Kaspersky’s latest India data provides an early indication of that change. The company’s anonymised Safe Kids data shows ChatGPT appearing among the top 20 apps used by children in India, while Android devices show a substantially higher share of detected adult-content activity than Windows devices.
Kaspersky and KidZania India presented those findings during a Media Family Day at the Kaspersky Cyber Investigation Centre at KidZania Delhi NCR in Noida on October 7.
The event itself is not the most important technology development. The more consequential story is what the data suggests about the changing nature of children’s digital behaviour—and what cybersecurity education now needs to teach.
AI Is Becoming Part of Children’s Digital Lives
According to Kaspersky’s Safe Kids data covering a recent six-month period in India, ChatGPT accounted for 2.41% of children’s app usage in the company’s Top 20 ranking.
It placed seventh, ahead of Zoom, Roblox and Telegram, while remaining far behind YouTube at 30.68%, WhatsApp at 29.14% and Instagram at 10.24%.
The ranking does not show that AI has displaced conventional children’s platforms. It shows something more subtle: a general-purpose AI chatbot has become significant enough to appear alongside some of the digital services children already use regularly.
Kaspersky says ChatGPT was the only AI chatbot in India’s Top 10.
The company’s wider global research provides additional context. Kaspersky’s 2026 analysis of anonymised Safe Kids data found AI tools dominating searches within its “Apps and tools” category. The company recorded searches for ChatGPT, Gemini, DeepSeek, Claude, Copilot, Grok and Character AI among children’s searches during March–May 2026.
That does not mean every child is using AI in the same way.
Some may use chatbots for homework. Others may seek quick answers, explanations, entertainment or conversation. The important point is that AI is increasingly becoming an interface through which children obtain information.
That changes the safety equation.
The New Risk Is Not Always a Malicious Link
Traditional cybersecurity education often focuses on recognisable threats.
Children learn not to share passwords, click suspicious links, respond to strangers or download unknown files.
Those behaviours remain essential. But AI introduces a different problem: the technology itself can produce an answer that looks credible without necessarily being correct.
Kaspersky General Manager for India Jaydeep Singh argues that children now need to learn to question AI responses as well as conventional online threats.
“It is no longer enough to tell a child not to click a suspicious link.”
The company’s position is that children should also learn to question what a chatbot tells them, recognise manipulative messages and understand when to seek help from an adult.
That is an important shift from cybersecurity awareness toward AI literacy.
UNICEF’s work on responsible AI for children points in the same direction. Its research identifies risks including AI-generated misinformation, privacy concerns, emotional dependency on companion chatbots and harmful synthetic content. It also calls for AI literacy, data protection, transparency and human oversight.
UNICEF’s Innocenti research similarly identifies AI literacy, digital literacy, critical thinking and online-safety awareness as complementary competencies children need in an AI-shaped environment.
The technology therefore changes the question children must learn to ask.
It is no longer simply:
“Is this link safe?”
It increasingly becomes:
“Can I trust what this technology is telling me?”
Smartphones Show a Different Exposure Pattern
Kaspersky’s India data also highlights a significant difference between Android and Windows devices.
On Windows, Internet, Communication & Media represented 43.44% of activity among the categories measured by Safe Kids, followed by Software, Audio, Video at 34.99%.
Android showed a similar pattern, with Internet, Communication & Media accounting for 39.86% and Software, Audio, Video for 34.89%.
The more striking difference appears in adult-content activity.
Kaspersky reports that adult content represented 5.71% of children’s activity on Android, compared with 1.91% on Windows.
The difference deserves attention, but it also needs careful interpretation.
These figures come from anonymised Kaspersky Safe Kids users. They should not be treated as a representative measurement of all Indian children.
Nor do they prove that smartphones are inherently more dangerous than computers.
They do demonstrate that the risk profile observed by the company’s technology differs materially between mobile and desktop environments.
Kaspersky suggests that smartphones may receive lighter supervision in households, but the supplied figures alone do not establish that this is the cause of the difference.
That distinction matters for responsible reporting.
Why Mobile Changes the Safety Challenge
Smartphones collapse several digital activities into one personal device.
A child can move from a messaging conversation to a video, a social platform, a browser, a game and an AI chatbot without changing devices.
The boundaries between categories therefore become increasingly difficult to maintain.
The same device can be a communication tool, entertainment platform, research assistant and social environment.
Kaspersky’s global Safe Kids analysis also found mobile devices accounting for a larger share of communication activity than Windows computers, while adult-content detections were higher on personal smartphones than computers in its global dataset.
That makes mobile safety more than a matter of installing a filter.
It becomes a question of whether children have enough understanding to make decisions when technology presents something unexpected.
From Cyber Investigation to Critical Thinking
This is where Kaspersky’s partnership with KidZania becomes relevant.
The Cyber Investigation Centre at KidZania Delhi NCR gives children aged six and above a simulated role as cyber investigators.
They work through cases involving phishing, identity theft and cyberbullying at dedicated workstations equipped with Kaspersky security software.
Each session lasts approximately 20 minutes and accommodates up to six children. Participants receive a personalised Cyber Investigator card after completing their mission.
The centre also operates at KidZania Mumbai.
This is not a newly announced programme. Kaspersky and KidZania launched Cyber Investigation Centres in Mumbai and Delhi NCR in April 2026. The original programme was designed as an immersive role-play experience in which children identify online threats, practise safe browsing and respond to simulated cybersecurity scenarios.
That history is important because the October family day represents an extension of an existing educational initiative, not the launch of a new cybersecurity technology.
Its relevance to the broader story is experiential learning.
Why Hands-On Learning Matters
Kaspersky’s Jaydeep Singh says the family event was intended to let parents see children learning cybersecurity through practice.
His argument is that when a child works through a simulated phishing investigation and then explains the experience to a parent, the lesson becomes more memorable than a conventional warning.
KidZania’s Chief Business Officer, Tarandeep Singh Sekhon, makes a similar case.
The company positions its Cyber Investigation Centre as an example of its broader role-play model, in which children learn by experiencing real-world professions rather than simply receiving instructions.
That approach has particular relevance to cybersecurity.
Threats evolve too quickly for children to memorise a fixed catalogue of dangers.
The objective should instead be to build habits of observation, questioning and escalation.
A child who understands why a phishing message is suspicious has a more transferable skill than one who simply memorises a particular example.
The same principle increasingly applies to AI.
Parental Controls Have a Limit
Kaspersky’s recommendations include keeping communication open, applying content filters and parental controls to smartphones, teaching children that AI can be wrong, helping them identify manipulation and periodically reviewing privacy and safety settings.
Those are sensible measures.
But they also expose a limitation of technology-led approaches to child safety.
A filter can potentially block a category of content.
It cannot teach a child why an AI answer may be unreliable.
A parental-control system can restrict access to an application.
It cannot necessarily teach a child why sharing a personal detail with an AI service may create a privacy concern.
Monitoring can identify behaviour.
It cannot replace judgement.
That means the next phase of child online safety will require both technical controls and behavioural literacy.
India Already Has a Broader Child-Safety Framework
The issue is also moving beyond individual technology companies.
The Indian government said in August 2026 that it was examining global developments in children’s online safety while reviewing policy and legal frameworks around digital well-being, privacy and emerging online risks.
The government pointed to the IT Act, IT Rules and Digital Personal Data Protection Act as part of India’s framework for digital safeguards. It also highlighted risks including harmful content, cyberbullying, online exploitation and digital dependency.
The government’s response also notes existing cyber-safety resources from the National Commission for Protection of Child Rights, digital education guidance and cybersecurity initiatives involving schools and educators.
That creates an important policy context for the Kaspersky findings.
The challenge is no longer confined to deciding which websites children should be allowed to visit.
It increasingly involves determining what skills children need to navigate systems that can generate information, personalise interactions and influence behaviour.
AI Literacy Must Become Part of Cyber Safety
The Kaspersky data should not be treated as evidence that AI is inherently harmful to children.
It is evidence of adoption.
That distinction is critical.
AI can help children learn, translate information, explore ideas and develop new skills. UNICEF’s research similarly recognises both the opportunities and risks associated with children’s growing use of AI.
The challenge is therefore not to keep children away from AI.
It is to ensure that they understand how to use it safely.
That requires several layers of education.
Children need to know that AI can be wrong.
They need to understand that not every chatbot interaction is private.
They need to recognise manipulative prompts and inappropriate requests.
They need to verify important information rather than treating an AI response as an authority.
And they need to know when an interaction should be brought to an adult.
These are not merely cybersecurity skills.
They are AI-era digital literacy skills.
The Bigger Shift: From Cyber Hygiene to Digital Resilience
Kaspersky’s family-day initiative offers a useful illustration of where child cybersecurity education is heading.
The company’s Cyber Investigation Centre teaches children to identify concrete threats such as phishing, identity theft and cyberbullying.
Its latest India data suggests that the environment around those threats is changing.
Children are increasingly using smartphones as their primary digital interface, while AI tools are becoming part of the same ecosystem.
The result is a broader definition of online safety.
Cyber hygiene teaches children how to avoid known threats.
Digital resilience prepares them to respond when the threat does not look familiar.
That distinction could become increasingly important as AI systems become embedded in search, education, messaging, entertainment and other services children already use.
The most valuable lesson may therefore not be “don’t click that link.”
It may be:
Pause. Question it. Verify it. And ask for help when you’re unsure.
That is a cybersecurity habit that can survive the next technology shift.
Editor’s Note
This article uses Kaspersky’s India Safe Kids figures and statements from the Kaspersky–KidZania event as company-reported information. The figures represent anonymised users of Kaspersky’s products and should not be interpreted as a representative survey of all Indian children. Independent context is drawn from Kaspersky‘s published global methodology and UNICEF’s work on children, AI literacy and online safety, as well as Indian government material on children’s digital protection.
The Cyber Investigation Centre was launched by Kaspersky and KidZania in Mumbai and Delhi NCR in April 2026; the October family day is therefore treated as an extension of an existing educational programme rather than a new technology launch. Analysis and interpretation are TechRecast’s.