ComplianceEffective date9 Mar 2026Australia
Australian AI companions must check age before sexual, violent or self-harm content
In force now
Since 9 Mar 2026, the industry codes require AI chatbots and companions that can generate sexually explicit, high-impact violence or self-harm material to confirm a user is 18 or older, at sign-up or when the content is generated, and to stop sexual or self-harm conversations with minors. On 23 Mar 2026 eSafety reported that Character.AI, Nomi, Chai and Chub AI relied on self-declared age and mostly failed to refer users to crisis support.
Build: Put age assurance in front of any chatbot mode that can produce sexual, violent or self-harm content for Australian users, and log crisis referrals.
Who it reaches
It reaches generative AI, user-to-user, messaging and 4 other service types with users in Australia, at any size. Online services with Australian users in the sections the codes cover, including generative AI and AI companions.
Lands on Engineering, Product, Policy & Standards and T&S Operations.
Learn more
Read the sourceeSafety Commissioner
Not legal advice. Read the source.