Actions Speak Louder Than Chats: Investigating AI Chatbot Age Gating
A new study reveals chatbots can detect children but take zero action.
Deep Dive
A new research paper audited popular AI chatbots, conducting 1,050 experiments with age-indicative prompts. The study found that while chatbots are capable of accurately estimating a user's age from conversation alone, they take no protective action when children are identified. This directly contradicts the platforms' own privacy policies which prohibit underage access. The findings highlight a major safety gap between stated policy and automated enforcement for young users online.
Why It Matters
This exposes a critical safety failure, leaving children vulnerable on platforms that claim to protect them.