AI puts children’s privacy at risk 

Engr. Md Mushfiqur Rahman is the Chief Technology Officer (CTO) of Standard Islami Bank PLC and Adviser of cyber crime awarness foundation. He is a technology and cybersecurity professional with expertise in digital transformation, information security and technology management in the banking sector. He regularly focuses on emerging technology risks, digital security and responsible technology adoption. In an interview with the Daily Observer’s Mizanur Rahman, Engr. Md. Mushfiqur Rahman discussed the risks of children's AI use, privacy and cybersecurity concerns, deepfakes, parental monitoring and the need for coordinated initiatives by the government, schools, parents and technology companies. We share the interview with our readers: 

DO: What are the biggest risks of allowing children to use ChatGPT or generative AI?
Mushfiq: The biggest risk is not AI itself, but unsupervised and inappropriate use of AI by children. Children may not yet have the maturity to distinguish between accurate information, misinformation, manipulation and fabricated content generated by AI.

There are several major risks, including exposure to inappropriate content, over-reliance on AI for schoolwork, manipulation through highly personalised interactions, privacy loss, cyberbullying, scams and exposure to deepfakes.

Children may also unknowingly disclose sensitive information about themselves, their parents, school, location, daily routines or family.

Another concern is AI becoming an authority figure in a child's mind. Children may accept an AI-generated answer simply because it sounds confident and convincing. Therefore, children need to learn that AI should be treated as an assistant, not as an unquestionable source of truth.

DO: What long-term negative impact could excessive dependence on AI have on critical thinking, creativity and problem-solving?Mushfiq:Excessive dependence on AI could create a generation that is very good at getting answers but weaker at developing answers themselves.

If a child routinely asks AI to write essays, solve mathematics, generate ideas, summarise books or complete assignments, the child may gradually lose opportunities to practise reasoning, experimentation, imagination, persistence and independent problem-solving.

This is particularly concerning during childhood because cognitive skills are still developing. Critical thinking is a skill that must be exercised; it cannot simply be outsourced to technology.

However, AI does not necessarily reduce intelligence or creativity. Used correctly, it can do the opposite. Children can use AI to ask “why”, explore alternative solutions, receive explanations, brainstorm ideas and challenge their own thinking.

The principle should therefore be: AI should extend a child's thinking-not replace the child's thinking.

DO: What privacy and cybersecurity risks arise when children share personal information, photos, voice recordings or family information?
Mushfiq: This is a serious concern because children may not understand the long-term consequences of sharing information online. Photos, voice recordings, school names, family information, locations, routines and private conversations can expand their digital footprint, creating risks of data misuse, identity theft, social engineering, impersonation, profiling and targeted scams. AI may also enable convincing synthetic images and cloned voices. Parents should teach children never to assume that “it’s only AI.” Children must never share passwords, OTPs, financial information, home addresses, precise locations, school security information, private family conversations or other sensitive personal data with AI systems.

DO: To what extent is AI increasing the risk of harassment, fraud and cyberbullying through deepfakes and voice cloning?
Previously, creating a convincing manipulated video, photograph or voice recording required considerable technical expertise. Today, sophisticated generative AI tools can make such manipulation accessible to ordinary users.

For children, this creates a particularly serious threat. A student's photograph can potentially be manipulated into humiliating or abusive content and circulated among classmates or on social media. Similarly, voice cloning can potentially be used in impersonation scams.

The psychological impact can be devastating because children may face humiliation, social isolation, threats or reputational damage before adults even realise what is happening.

DO: Should children be completely restricted from AI, or should they use it safely and under control? What parental monitoring is necessary?
Mushfiq: I don’t believe complete prohibition is the best long-term strategy, as AI is becoming essential in education, work and society. Instead, children should use AI responsibly, appropriately and under supervision. Parents should guide rather than spy, know the tools used, set rules and discuss data safety.

Younger children need stronger supervision, while older children can gradually gain trust and accountability.

I would recommend five basic rules: Don't share sensitive personal information. Don't automatically believe AI-generated information-verify important facts.

 Don't use AI to bully, impersonate or deceive anyone. Don't allow AI to replace learning and independent thinking. Tell a trusted adult immediately if AI generates something frightening, sexual, threatening or abusive.

Parents should also maintain appropriate technical controls such as device-level parental controls, safe browsing, privacy settings, strong authentication and updated software.

DO: What coordinated initiatives should government, educational institutions and parents undertake?
Mushfiq: This cannot be solved by parents alone; government, schools, technology companies and families must work together. The government should establish child-AI safety, data-protection and age-appropriate AI standards, strengthen reporting of AI-enabled abuse and deepfakes, and ensure strong data security. 

Schools should introduce AI literacy and digital citizenship, teaching students to identify misinformation/deepfakes, protect data, verify AI content and use AI ethically. Clear policies should guide AI-assisted assignments without outsourcing learning. Parents should encourage children to discuss online experiences safely. AI companies should adopt child-safety-by-design, including age safeguards, privacy protection, abuse reporting, safety filters and transparency on children’s data.

DO: What is your overall message to parents and policymakers?
Mushfiq: We should not raise children who are afraid of AI; we should raise children who understand AI.

The solution is not to ban technology, but to combine AI literacy, parental supervision, strong privacy protection, cybersecurity education and responsible use. AI should be a tool that strengthens a child's curiosity, creativity and critical thinking-not a substitute for them.

From a cybersecurity perspective, the most important digital skill for the next generation will not simply be knowing how to use AI. It will be knowing when to trust it, when to question it, what information not to give it and when to ask a human for help.