Product assessment / Child safety · Safety & wellbeing
Common Sense Media reports testing over 4,000 prompts and identifies gaps in parental alerts, crisis referrals and relational boundaries, alongside some functioning safeguards. Its assessment includes an editorial note about OpenAI’s approximately three-hour alert-activation requirement; some tested accounts were linked for shorter periods, others longer.
Why it matters
High-priority review of product-level safeguards and expectations around parental notifications.
Read with care
This is the institute’s assessment, not a regulatory finding or a study of real children’s outcomes. US test accounts; model changes confound before/after comparisons; inter-rater agreement was not calculated. Four alerts occurred in the broader battery, so do not say alerts never work. Activation dispute remains unresolved. Institute discloses industry funding including OpenAI Foundation. HTML methods, findings and editorial note screened; raw logs not independently reproduced.
What was checked
Source record and stated publication date checked; see reading cautions for the extent and limits of the review. Checked 2026-10-12.