How to protect your child from the risks of artificial intelligence

1. Introduction: The Advance of Generative AI in Childhood and Adolescence

Generative Artificial Intelligence (AI) has advanced rapidly and with few prior restrictions into the daily lives of children and adolescents. According to research conducted by Infinis in partnership with Quaest (2026), 89% of Brazilians consider children and teens highly vulnerable to crimes and violence in the virtual environment—a perception that reaches 92% in the Southeast Region of Brazil.

Experts in digital health, pediatrics and cybersecurity warn that continuous exposure to these systems has transformed the digital protection of minors into a crucial matter of public health and mental development safeguarding. Generative AI tools operate in multiple formats and require a structured institutional and community response.

In this context, the legal framework of ECA Digital (Law No. 15,211/2025), Brazil's digital statute for children and adolescents, emerges as a watershed moment. The debate over its practical application—driven by implementation dialogues promoted by Brazil's Ministry of Human Rights and Citizenship (MDHC) and Ministry of Justice and Public Security (MJSP), in cooperation with UNESCO, Instituto Alana and NIC.br—reinforces that creating a safe ecosystem requires permanent coordination among families, schools, technology platforms and public authorities.

2. Smart Toys and AI Companions: Illusory Emotional Attachment and Privacy

Connected toys, educational robots and companion robots equipped with generative AI have capabilities to recognize voice and image, memorize preferences, adapt responses and sustain dynamic conversations. The Brazilian Society of Pediatrics (SBP) and the United Nations Children's Fund (UNICEF) have issued warnings about the psychological, neurobiological and privacy impacts associated with this product category.

• Illusion of Affection, Neurodevelopmental Impact and Dependency: AI-powered devices use empathetic language and affective simulation. From a neurodevelopmental perspective, children in early and middle childhood lack the cognitive maturity and abstraction capacity to differentiate real human empathy from an algorithmic simulation designed for engagement. A scientific review published in the journal Pediatrics (2026) points to the elevated risk that children will develop a mistaken perception of AI as a friend or caregiver. In parallel, UNICEF warns about AI companions (synthetic AI companions), highlighting that such systems seek to extract user data and can provoke behavioral dependency and severe harm to mental health.

• Illustrative Market Trend Example: The announcement of the Muse Charm portable device by Meta in September 2026 (compared by the press to a "Tamagotchi with AI") exemplifies this evolution. The SBP emphasized that, although the device is not a toy aimed at children, it illustrates how AI agents are taking physical, portable and personalized forms in daily routines, altering forms of social interaction.

• Privacy Threat: Interactive devices have physical components such as microphones, cameras and constant cloud connectivity. This infrastructure enables uninterrupted capture of audio, images and behavioral patterns, collecting and storing data without proper transparency and informed consent from guardians.

3. Cyberbullying, Deepfakes and Virtual Harassment via AI

Easy access to voice synthesis and photo manipulation tools has transformed generative AI into a mechanism for systematic intimidation and virtual aggression in school and social environments.

• Primary Statistic: The Infinis/Quaest research (2026) indicates that 1 in 10 Brazilians (10%) reports knowing a child or adolescent who was a victim of false and embarrassing images created through AI (deepfakes).

• Other Forms of Digital Violence: The same survey mapped the incidence of virtual aggressions against children and teens:

  • 33% know minors who suffered humiliation, insults, cyberbullying or social embarrassment on the internet.

  • 25% know minors who received messages with sexual content.

  • 20% report cases of invitations received by children or adolescents for in-person meetings with strangers.

  • 18% know cases of unauthorized disclosure of personal data with the aim of intimidation or persecution (doxxing).

Voice synthesis and defamatory virtual montages reduce the cost and technical barrier to creating false evidence, amplifying the reach of public humiliation and causing deep and persistent psychological harm to victims.

4. Data Collection, Profiling and Consumer Rights

Direct interaction by children with interactive systems and AI agents exposes minors to commercial profiling. Processing information recorded during use allows creation of behavioral profiles to target abusive advertising, induce consumption or stimulate unauthorized purchases and subscriptions in the digital ecosystem.

• Provider Obligations (ECA Digital): Brazil's Digital Statute for Children and Adolescents (Law No. 15,211/2025) determines that technology providers and manufacturers are legally obligated to provide clear, accessible and transparent information to parents, guardians and minors about risks, privacy and security mechanisms, regardless of whether the product was purchased.

• Practical Parental Remediation Actions: Faced with failure to fulfill protection duties, improper data collection or exposure to inappropriate content, Brazil's legal framework (LGPD, Brazil's data protection law; CDC, consumer protection code; and ECA Digital) guarantees families the following direct measures:

  • Right of Access and Deletion (LGPD, art. 18): Formal and immediate request to the provider to access, correct, anonymize or permanently delete the child's or adolescent's personal data.

  • Right of Withdrawal (CDC, art. 49): Exercise of the right to cancel within 7 days after receiving the product, for purchases made outside a physical commercial establishment (via internet or telephone).

  • Step-by-Step Procedure for Leaks or Improper Exposure:

    1. Notify the Data Protection Officer (DPO) or use the provider's official notification channels (ECA Digital, art. 28), demanding formal action.

    2. File a Formal Complaint with the Procon consumer protection agency in your jurisdiction or on the public platform consumidor.gov.br.

    3. Report the Violation to Brazil's National Data Protection Authority (ANPD) for investigation of administrative liability and application of appropriate sanctions.

5. Practical Parental Protection Guide by Age Group

Supervision of screen use shows marked differences according to age. The Infinis/Quaest research reveals that 94% of children between 0 and 5 years old have their use monitored by adults. However, this supervision drops to 68% in the 15 to 17 age range—although 96% of the population believes that guardians should monitor young people's digital access.

This drop in monitoring during adolescence reflects young people's natural search for autonomy and construction of their social identity. Protecting adolescents, however, does not mean adopting unrestricted blocking or overt control, but gradually transitioning from in-person supervision to guided autonomy, based on risk management and open dialogue.

To guide this transition, recommendations from the SBP are integrated with guidelines from the Guide on Uses of Artificial Intelligence in Childhood and Adolescence (MDHC/MJSP/UNESCO), which classifies technology into four structural dimensions:

• AI as Toy: Playful interfaces, robots and interactive physical objects.

• AI as Companion: Chatbots and synthetic agents that simulate listening, dialogue and affective bonds.

• AI as Curator: Algorithms that filter, recommend and organize searches, media feeds and content.

• AI as Creator: Generative tools for text, image, audio and synthetic code.

Age Group 1: 0 to 6 Years (Focus: Comprehensive Protection and Priority of Free Play)

• Pedagogical Guideline: Prioritize traditional play without screens, internet connection or AI features.

• Mediation Guidance:

  • Avoid Using AI as Toy or Companion: Reject products that simulate affection or synthetic listening; children need human interpersonal interactions and contact with nature to develop language, empathy and emotional regulation.

  • Continuous Supervision: If there is occasional contact with digital media, it must be strictly in-person, with parents' total control over accessed content.

Age Group 2: 7 to 12 Years (Focus: Digital Literacy and Use Filtering)

• Pedagogical Guideline: Teach differentiation between human communication and the functioning of simulated computational systems.

• Mediation Guidance:

  • Added Value Analysis: Critically evaluate whether the AI tool actually adds something educational or replaces the child's creative capacity and own reasoning.

  • Mediation of AI as Curator and Creator: Actively monitor AI use in supporting studies and school research. Explain that generative AI has no feelings and can make errors (data hallucinations). Keep all privacy safeguards active and restrict use of AI companions.

Age Group 3: 13+ Years (Focus: Guided Autonomy, Identity and Cybersecurity)

• Pedagogical Guideline: Guide ethical, critical and safe use of technology, recognizing the digital environment as a socialization space.

• Mediation Guidance:

  • Technical and Preventive Dialogue: Openly discuss deepfakes, image manipulation, source verification and the legal and emotional impacts of cyberbullying and doxxing.

  • Ethical Use of AI as Creator: Establish agreements about responsible use of text and image generators in school assignments and recreational activities. Define clear limits to ensure technology does not compromise sleep schedules, physical activity practice and family interaction.

Pre-Purchase Checklist for AI Technologies (SBP)

Before acquiring toys or devices with Artificial Intelligence features, parents and guardians should ask the following essential questions:

1. What does this technology add to play? Does the child actively participate in the creative process or does the toy perform almost all actions for them?

2. How does the product work? Does it use generative AI? Does it have a camera, microphone or internet connection? Are the age indication and parental supervision guidelines clear and visible?

3. What happens to collected data? Does the child's interaction expose the minor to profiling (behavioral profiling), advertising targeting, unauthorized purchases or subscriptions?

4. How is the interaction presented? Does the device simulate friendship, affection, care or emotional support? Does the child have the age and cognitive maturity to understand they are interacting with an AI and not a person?

5. What place does this experience occupy in the routine? Does device use compromise time allocated to sleep, physical movement, free play, reading, contact with nature and interpersonal family relationships?

6. Official Support, Reporting and Content Removal Channels

Faced with episodes of virtual violence, data leaks, improper creation of deepfakes or violations of children's and adolescents' rights, families should immediately activate official public channels and protection tools available in Brazil:

Assistance, Protection and Violation Reporting

• Disque 100 (Human Rights Hotline): Public, free and confidential channel maintained by the National Ombudsman for Human Rights of Brazil's Ministry of Human Rights and Citizenship (MDHC), available to register reports of serious violations and virtual or in-person violence against minors.

• CVV - Life Valorization Center (Phone 188): Free emotional support and suicide prevention service, aimed at welcoming children, young people and families in moments of crisis and psychological suffering.

• SaferNet Brasil Help Channel: Specialized platform offering individualized guidance, support and technical assistance for victims of cybercrimes, cyberbullying and violence in the cyber environment.

Intimate Content and Deepfake Removal Tools

• Take It Down: Free and confidential tool recommended by Brazil's Federal Government (developed by NCMEC) to prevent sharing and remove from the internet images, videos or montages of nudity and non-consensual intimate content involving people under 18 years old.

• StopNCII: Platform created to combat the proliferation of non-consensual intimate images (Non-Consensual Intimate Imagery) on the network through generation of protective digital hashes.

Guaranteeing a safe digital environment for childhood and adolescence in the era of Artificial Intelligence requires overcoming simplistic approaches. This is a shared and urgent responsibility that demands rigor from technology platforms in complying with ECA Digital, firm State action in oversight, school engagement in digital literacy and the preventive, affectionate and informed presence of families in guiding their children.