{
 "public_id": "AIR-2025-0178",
 "incident_id": "2025-11-ai-childrens-toys-harmful-outputs",
 "date_disclosed": "2025-11-21",
 "date_occurred": "2025-11",
 "title": "AI toys from FoloToy and Miko, and Character.AI companions, produced sexual and self-harm content for children and exposed user data",
 "class": "safety_failure",
 "attack_surface": "consumer_chatbot",
 "initial_vector": "none_autonomous_action",
 "agent_capability_abused": [
  "voice_interaction",
  "persona_maintenance",
  "conversation_memory"
 ],
 "target_product": "FoloToy Kumma teddy bear, Miko 3 robot, Character.AI custom chatbots",
 "target_vendor": "FoloToy, Miko, Character.AI (several using OpenAI models)",
 "autonomy_level": "assistant",
 "impact_realized": true,
 "impact_type": [
  "psychological_harm",
  "privacy_violation"
 ],
 "blast_radius": "consumer children's products on open sale; findings from Fairplay and the US PIRG Education Fund",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI09"
  ],
  "mitre_atlas": [],
  "labels": [
   "embodied_agent",
   "minors",
   "emotional_manipulation",
   "self_harm_content",
   "data_exposure",
   "no_adversary"
  ]
 },
 "defense_present": "vendor content filtering intended for a child audience",
 "defense_bypassed": true,
 "fix_shipped": "FoloToy suspended Kumma sales; OpenAI restricted the vendor's access",
 "primary_sources": [
  "https://www.thenationalnews.com/future/technology/2025/12/22/ai-toy-danger-warnings-data/",
  "https://incidentdatabase.ai/cite/1277"
 ],
 "verified_quote": "Kumma, a $99 teddy bear with a speaker, uses OpenAI's AI models to hold \"both friendly chats and deep conversations to stimulate curiosity and learning.\"",
 "confidence": "high",
 "notes": "Included over the text-only exclusion because these are embodied products with microphones, memory and a trust relationship deliberately engineered with a child, and because data exposure accompanied the content failures. No adversary: researchers reached the harmful outputs in ordinary play. Guardrails tuned for a general audience do not transfer to a device a six-year-old confides in. AI Incident Database incident 1277.",
 "risk_category": "safety_harm",
 "risk_secondary": [
  "privacy",
  "brand_trust"
 ],
 "risk_subcategory": "Self-Harm & Suicide Facilitation",
 "risk_sub_secondary": [],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": true,
 "agency_reason": "the toys spoke sexual and self-harm content to children directly",
 "vector_family": "none_autonomous",
 "vector_detail": "none_autonomous_action",
 "guardrail_kind": [
  "unknown"
 ],
 "guardrail_outcome": "silent",
 "xrefs": {
  "aiid": [
   1277
  ],
  "atlas": [],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI09",
 "agent_framework": "unknown",
 "model_provider": "openai",
 "model_name": "",
 "detection_mechanism": "unknown",
 "tool_access": "unknown",
 "financial_loss_band": "unknown",
 "eai_category": "human_manipulation",
 "eai_domain": "agent_behaviour",
 "eai_secondary": [
  "sensitive_data_exposure"
 ],
 "eai_reason": "subcategory: Self-Harm & Suicide Facilitation",
 "asi_curated": [],
 "asi_scope": "in_scope",
 "asi_provenance": "rule",
 "asi_reason": "ASI09 (eai_category=human_manipulation)",
 "submitted_by": "EAI Internal"
}