What happened
On 28 October 2026, the Online Safety Amendment (Digital Duty of Care) Bill 2026 passed the Australian House of Representatives with amendments and is now before the Senate. The bill would require the person responsible for an online service to take 'appropriate and effective' action to protect users, including mitigating risks identified in annual risk assessments, replacing existing industry codes and standards with a single regime enforced by the eSafety Commissioner. It applies to websites, games, messaging apps and AI chatbots, adds a child-specific duty covering harmful material, features and design (including content presented during a chatbot conversation), gives the commissioner powers to issue removal notices for apps/websites designed to generate fake nude material, and proposes a choice of primary social-media feed for users over 16 (algorithmic vs followed accounts). The headline civil penalty is A$109.2 million, with a proposed 12-month transition before the duty commences; a Senate committee must report by 20 November 2026.
Why it matters
This is a major shift in Australian online-safety regulation that explicitly covers AI chatbots and AI-generated content, and would move from reactive complaint handling to proactive risk-mitigation duties backed by a nine-figure penalty. For AI chatbot and generative-AI providers, it introduces obligations to address harms arising from features and tools (not just individual posts), child-specific safety duties, and the risk of removal notices for AI-generated fake-nude services. Its passage would materially change the operating bar for AI services in Australia.
Action needed
AI chatbot and online-service providers operating in Australia should assess how their service exposes users to covered harms, map risks arising from features/tools, and prepare for the child-safety duty and the proposed feed-choice requirement, given the bill's passage through the House and the imminent Senate committee report (due 20 November 2026).