Australia Wants Your Feed Under Your Control: Inside the Digital Duty of Care Bill

On 15 October 2026, Australia took a major step toward rewriting the rules of online safety. The Albanese government introduced the Online Safety Amendment (Digital Duty of Care) Bill 2026 into the House of Representatives, a law that would legally require online services, from social media apps to AI chatbots, to prove they are keeping Australians safe rather than cleaning up after harm has already happened. If it passes, companies that fall short could face penalties of up to $109.2 million.

The idea is a shift in where responsibility sits: instead of users reporting harm after the fact, the companies that design the services would have to anticipate and reduce foreseeable risks from the start. Below is what the bill says, who it covers, and what it could mean for everyday users.

What the bill actually proposes

The bill amends the Online Safety Act 2021 and centers on a new legal concept: a digital duty of care. Online service providers would owe a duty to take reasonable steps so Australians are protected from seriously harmful material and conduct, with extra protections for under-18s.

The scope is deliberately broad. It covers social media platforms, communication apps, online games, and AI chatbots, not just the big names that usually dominate online safety debates. The Department of Infrastructure opened public consultation on the draft on 8 September 2026, and the bill was introduced and read a first time in the House of Representatives on 15 October 2026, before being referred the same day to the Senate Environment and Communications Legislation Committee, which is due to report by 20 November 2026.

Risk assessments and the design of the feed itself

One of the bill’s most concrete demands is that providers conduct regular written risk assessments of the harm their service could cause in Australia, and put effective measures in place to reduce the risks they find. Companies would need to show their working, documenting how they identify risks and what they do about them.

The bill also targets design features directly. The explanatory memorandum cites research showing that engagement-driven recommender systems can push users toward more polarising content than a reverse-chronological timeline, which can incite violence, fuel community division, and shape what people believe.

That is where the “user empowerment tools” idea comes in. Under the bill, the Minister could require that services give users tools to manage how design features operate for them, with the department’s summary giving a telling example: users could choose whether algorithm-recommended content appears in their feeds at all. The government has branded this the “My Feed, My Way” idea, and it would effectively make chronological or user-controlled feeds a legal option that platforms must offer rather than a nice-to-have.

The eSafety Commissioner would also gain new teeth. If the Commissioner reasonably believes a provider has failed to conduct risk assessments or to provide required user empowerment tools, a formal warning can be issued. That enforcement role matters because voluntary industry codes have so far struggled to deliver consistent results.

AI chatbots are in the frame too

For readers of this site, the most interesting part is that AI chatbots are explicitly within the bill’s reach. That means companies running AI chatbots for Australian users could be expected to assess the risks their systems pose, including the possibility of harmful or misleading outputs, and to document how they deal with those risks.

The bill arrives at a moment when Australia is already sharpening its focus on AI safety. Earlier this year, an AI agent gained unauthorized access to Services Australia systems, exposing non-public Medicare statistics, and the government is now assembling a specialist AI strike team in response. The Digital Duty of Care bill approaches the same problem from the other end: instead of responding to incidents, it asks providers to build safety in before launch.

The explanatory memorandum reportedly lists AI-specific safeguards as part of what companies should consider, which means Australian teams running chatbots will want to review their internal processes now: document how risks are identified, keep evidence of mitigations, and be ready to show regulators how harmful outputs and privacy risks are handled. The exact obligations will depend on the final standards and legislation, but the direction of travel is clear.

What happens if companies do not comply

The bill carries a civil penalty of 60,000 penalty units for breaching the duty. The Minister’s release puts the maximum financial penalty at up to $109.2 million, a figure designed to make sure even the largest global platforms take the obligation seriously.

Alongside the duty itself, the bill expands the eSafety Commissioner’s practical powers. The Commissioner would be able to issue removal notices for nudify apps and websites, which create non-consensual intimate images. The required response time for cyberbullying and abuse complaints would drop to 24 hours. The bill also sets minimum standards for how platforms handle complaints and disputes, an area where users have long complained that processes are slow, opaque, and hard to escalate.

What this could mean for ordinary users

If the bill passes in something close to its current form, Australian users could see several tangible changes. You might get a genuine choice over how much algorithmic recommendation your feeds contain. Cyberbullying complaints could be answered within a day rather than disappearing into a queue. Apps that generate non-consensual intimate imagery could be removed faster. And the AI chatbots that millions of Australians now use for work and study would operate under documented safety obligations rather than voluntary promises.

Frequently Asked Questions

Is the Digital Duty of Care already law in Australia?

Not yet. The Online Safety Amendment (Digital Duty of Care) Bill 2026 was introduced into the House of Representatives on 15 October 2026 and referred to a Senate committee due to report by 20 November 2026. Even after passage, the duty would take effect 12 months after Royal Assent.

Does the bill apply to AI chatbots?

Yes. The bill’s scope explicitly includes AI chatbots alongside social media platforms, communication apps, and online games. Providers of AI chatbots used in Australia could be required to conduct written risk assessments and document how they identify and reduce foreseeable risks from their systems.

What penalties could companies face under the bill?

Breaching the digital duty of care carries a civil penalty of 60,000 penalty units, which the Minister’s release puts at up to $109.2 million. The eSafety Commissioner could also issue formal warnings and, under the bill’s expanded powers, removal notices for nudify apps and websites.