Today the House passed the Online Safety Amendment (Digital Duty of Care) Bill 2026, and I am so pleased that it did. I have called for a digital duty of care for a long time, particularly for Australians to have a real choice over the algorithms that decide what we see, because at the heart of so many harms we see online is an algorithm. These platforms are profiting from our attention. Every feature, from endless scroll and autoplay to constant notifications and like counts, is designed to keep us scrolling, and the algorithm learns exactly what keeps each of us individually hooked. Too often, what keeps us hooked is what makes us angry. A 2021 study of 2.7 million posts on Facebook and Twitter found that every additional word attacking the political 'other side' raised the odds of a post being shared by 67 per cent. Outrage is not a bug in these systems; it is a feature.

The consequences are real. In his annual threat assessment in June, ASIO director-general Mike Burgess warned that social media is fuelling Australia's rising terrorist threat—a terrorist threat that my community in Wentworth, in Bondi, has experienced directly—and that algorithms can take someone from non-violent content to ISIS propaganda with only a few clicks. Fewer than 17 per cent of the minors ASIO investigated were born overseas. These are Australian kids being radicalised in Australian bedrooms. ASIO's evidence is that just 10 minutes engaging with incel material can lead algorithms to recommend increasingly violent, misogynistic content. Dublin City University researchers found accounts set up to mimic teenage boys were served antifeminist content within 23 minutes. University of Melbourne research found misogynistic attitudes were the strongest predictor of support for violent extremism, with 19 per cent of 18- to 29-year-olds agreeing that violence could be legitimate to 'resist feminism'.

At a time of national crisis in domestic and family violence, we cannot ignore a machine feeding young men a steady diet of contempt for women. That is why I've championed Teach Us Consent's Fix Our Feeds campaign since the end of last year. Its ask is simple: put the choice back in the hands of the users. Each of us should be able to decide whether we want the algorithm, a feed of people we know, or something depersonalised. Some people want to switch off the algorithm for good. Others just want a break. Either way, the choice should be theirs, and I am proud that this choice is now part of the digital duty of care. I want to thank Chanel Contos and the team at Teach Us Consent, whose advocacy has been monumental, and the parents, young people and experts who told me over and over how harmful these algorithms can be.

I also want to be clear about what the bill is not. It is not a misinformation law. I voted against the mis- and disinformation laws that were brought in before, so I should know, because I take this very seriously. It does not ask the government or the platforms to decide what is true. It does not give the government power to censor anyone. It is for the parliament, not a minister, to decide any new category of harm. This is, first and foremost, a safety bill. If anything, it hands power back to the users from a handful of global companies that have had far too much power over us for far too long.

There is more to do. I will keep on pushing for an external dispute resolution scheme, stronger penalties, a right to redress and tools that are easy to find and easy to use. But today is a significant step. If you build it and you profit from it, you are responsible for making it safe. That is what it's like in a building, that is what it's like in clothing, that is what it's like in food, and that's what it should be like in social media as well. For the first time, Australians will have a real say over their feeds, and I think this is a good thing.