The Commonwealth Government introduced the Online Safety Amendment (Digital Duty of Care) Bill 2026 on 15 October, proposing proactive safety obligations for online services, including AI chatbots, games and apps.
The bill would require providers to address seriously harmful material and conduct, as well as risks arising from their services’ features and tools. Additional protections would apply to users under 18. The government says non-compliance could attract penalties of up to A$109.2 million. These are proposed obligations: the bill has not become law.
The package would also strengthen the eSafety Commissioner’s powers against AI “nudify” services. Separately, consultation has opened on proposed rules that would let social-media users choose between algorithmically recommended feeds and following-only content.
For AI providers and their lawyers, the consequential shift is toward scrutiny of service design—not simply removing individual harmful outputs after a complaint.
If enacted, the framework could make foreseeable risks from conversational features, engagement mechanisms and safeguards central compliance questions. Advisers will need to examine which services it covers, how risks must be assessed, and what evidence shows protective measures actually work. A warning in terms of use may be relevant, but it would not necessarily answer questions about an unsafe product feature.
Sources: ministerial announcement; feed-choice consultation.
Leave a Reply