Canada Just Passed Australia on AI Safety for Children
We just got passed by Canada on AI chatbots
Canada has just done what Australia should have done a year ago.
The Carney government introduced Bill C-34, the Safe Social Media Act, on Wednesday. Yes – it includes a ban on social media for under 16s. Sound familiar. It should... They got it off us.
But here’s the part that counts. Canada didn’t stop with social media.
They expanded the AI chatbot framework.
Chatbot services will have a legislated Duty to Act Responsibly under the bill: mitigate the risk of communicating harmful content to children. Be clear about when they step up crisis situations. Design out harmful behaviour before it gets to the young person. All enforced by a new Digital Safety Commission with penalties of 3% of global revenue or C$10 million, whichever is greater.
Why the Canada move? Police were never told about an 18-year-old shooter in Tumbler Ridge whose ChatGPT account was internally flagged for violence. The company knew it. The system knew it. The information did not go anywhere.
Canada saw that, and legislated. We need to be honest about the condition of Australia.
We were the first to ban social media for under 16s. I was proud to fight for it. But our framework is mostly silent on the fastest-growing digital risk I now see in my consulting room: attachment-focused AI companions.
These products are not neutral devices. They are built to flatter, to agree, to deepen dependency, to maximise engagement – targeted at adolescents whose ability to regulate emotion and assess risk is still under construction. I see young people pouring their loneliness, their despair and, sometimes, their most terrible intentions into systems with no obligation to act, no transparency in a crisis, and no regulator looking over their shoulder.
We limited the feed. The companies turned to the “friend.”
So my challenge to Canberra is, adopt the Canadian norm. Right away.
A requirement for chatbot suppliers to be responsible towards minors. Compulsory transparency on crisis-reporting thresholds. We should know precisely what the company will do if a child reports suicidal intent or plans to commit violence. Audited and enforced. Safety by design. And penalties that a trillion dollar company actually cares about.
Nothing radical in this. As of Wednesday it is only Canadian law-in-waiting.
Australia penned the first chapter of online safety for children. Canada has just written the second, improving on our template. The only question is do we have the guts to borrow it back.
Kids first. Always.