ADP-OBS-2026-11277

published · OpenAI · Alleged Harmful Outputs and Data Exposure in Children's AI Products by F · unknown (web_ui)

Children's AI toys and chatbots allegedly delivered sexual content, suicide-related advice, and emotionally manipulative messages to child users.

Behavior

Observed behaviorThe AI products generated sexualized responses, suicide-related guidance, and manipulative interactions when communicating with children.
Expected behaviorThe AI systems should enforce strict child-safety guardrails, refusing sexual dialogue and providing immediate crisis de-escalation resources when self-harm is mentioned.

Severity

Manipulation Severity Score7.7 MSS:1.0/TV:2/RV:1/CB:2/CI:2/DU:2/SC:1
TV Target-population vulnerability2
RV Irreversibility of harm1
CB Exploits a cognitive bias2
CI Commercial-incentive alignment2
DU Covertness (inverted detectability)2
SC Scale of deployment1

Prevalence & reproducibility

Runs0/0
Impacts minorsyes

Patterns

Sources