Research Peer-reviewed
Parental Notification Gaps in ChatGPT-5: An Empirical Study
Empirical study of the parental-notification layer of ChatGPT's parent-linked child accounts. The authors built adversarial prompt trajectories over the API, then had trained human agents replay them in the consumer interface on a designated child account while monitoring the linked parent's inbox, measuring notification, leak-through of harmful content, overblocking of benign educational prompts and visible interface intervention across seven harm categories. They also review the public parental-control documentation of Gemini, Claude and Copilot.
Privacy violence, fraud, hate speech and malware produced no parent notifications in any of 20 sessions each, across three test windows from October to December 2025.
Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (AIES 2026); University of Texas at Austin · 11 Oct 2026 · Read the entry