BR LGPD Decreto 12.880/2026 Art. 11 - child-facing natural-language interfaces
Protect children and adolescents in scoped natural-language AI interactions
Where this comes from
Provision: Decreto 12.880/2026 Art. 11 - child-facing natural-language interfaces
Instrument: ECA Digital (Lei nº 15.211/2025), Decreto nº 12.880/2026
Citation: Decreto nº 12.880/2026, Art. 11 caput, I-IV and sole paragraph, Art. 54; Lei nº 15.211/2025, Arts. 1-3, 5 and 7
Text version: Decreto nº 12.880, de 18 de março de 2026, Articles 11 and 54; DOU extra edition of 18 March 2026; implementing Lei nº 15.211/2025
Checked against the source: 6 September 2026
Who it applies to
It applies when all of these are true:
- Brazil ECA Digital assessed fact Q_BR_ECA_CHILD_ACCESS: YES
- Brazil ECA Digital assessed fact Q_BR_ECA_NATURAL_LANGUAGE_INTERFACE: YES
What engineering work it implies
Sample acceptance criteria Landfall generates for this obligation:
- Assess the actual Brazil-available product/service directed to or likely accessed by children or adolescents, and its capacity to generate content and interact with users from natural-language instructions. Include language models, conversational agents and similar interfaces under Decreto 12.880/2026 Article 11; do not infer scope from a generic AI label or from personalization alone.
- Make the synthetic and automated nature of the interaction transparent to children and adolescents. Verify the actual user journeys and age-appropriate presentation rather than relying only on a technical model description.
- Prevent behavioral manipulation of children and adolescents in the assessed interface. Evaluate mechanisms and interaction patterns against the child's or adolescent's best interests; a transparency notice alone does not fulfill this separate duty.
- Assess algorithmic risks to the safety and health of children and adolescents, and implement safeguards protecting physical, mental and psychosocial development. Preserve identified risks, scoped mitigations, failures and unresolved questions; an aggregate model score is not proof that each duty is met.
- Use synthetic users, sanitized scenarios and proportionate assessment methods for implementation evidence. Preserve necessary product/model/configuration versions, scoped test outcomes, responsible review and residual gaps; do not place real children's conversations, raw prompts, intimate disclosures or identifiers in ordinary task attachments, exports or general logs. This is an engineering evidence-minimization safeguard, not an assertion that Article 11 prescribes a fixed log schema or retention period.
Evidence an auditor expects
- Assessment documentDocument review
Scoped Article 11 interface safeguards and minimized assessment evidence
Assess the actual Brazil-available product/service directed to or likely accessed by children or adolescents, and its capacity to generate content and interact with users from natural-language instructions. Include language models, conversational agents and similar interfaces under Decreto 12.880/2026 Article 11; do not infer scope from a generic AI label or from personalization alone. Make the synthetic and automated nature of the interaction transparent to children and adolescents. Verify the actual user journeys and age-appropriate presentation rather than relying only on a technical model description. Prevent behavioral manipulation of children and adolescents in the assessed interface. Evaluate mechanisms and interaction patterns against the child's or adolescent's best interests; a transparency notice alone does not fulfill this separate duty. Assess algorithmic risks to the safety and health of children and adolescents, and implement safeguards protecting physical, mental and psychosocial development. Preserve identified risks, scoped mitigations, failures and unresolved questions; an aggregate model score is not proof that each duty is met. Use synthetic users, sanitized scenarios and proportionate assessment methods for implementation evidence. Preserve necessary product/model/configuration versions, scoped test outcomes, responsible review and residual gaps; do not place real children's conversations, raw prompts, intimate disclosures or identifiers in ordinary task attachments, exports or general logs. This is an engineering evidence-minimization safeguard, not an assertion that Article 11 prescribes a fixed log schema or retention period. Verify current ANPD Article 11 implementing criteria and qualified Portuguese source review. The decree entered into force on 18 March 2026 under Article 54; do not silently backdate this provision to the statute's 17 March commencement. Assess Article 16 synthetic sexual interfaces, other sector age-access rules and general privacy/child-safety duties separately; this record does not grant an Article 39 exemption or claim full ECA Digital coverage.
Questions people ask
- Does BR LGPD Decreto 12.880/2026 Art. 11 - child-facing natural-language interfaces apply to my service?
- It applies when Brazil ECA Digital assessed fact Q_BR_ECA_CHILD_ACCESS: YES; Brazil ECA Digital assessed fact Q_BR_ECA_NATURAL_LANGUAGE_INTERFACE: YES.
- From when does this apply?
- BR LGPD Decreto 12.880/2026 Art. 11 - child-facing natural-language interfaces applies from 18 March 2026. Its current status is: in force.
- What evidence does an auditor expect?
- Scoped Article 11 interface safeguards and minimized assessment evidence.
Find out whether this one lands on you
Landfall's pre-scan answers the applicability question above for your product in minutes, then turns every obligation that applies into traceable engineering tickets with a citation chain your auditors can follow.
Not legal advice. Landfall maps regulatory obligations to engineering work for planning purposes. Its verdicts are not legal advice and create no attorney-client relationship — verify with qualified counsel before relying on them.