The Australian government is preparing a draft law that would force social media platforms to give users the power to switch off the algorithmic recommendation systems. The algorithm controls what appears in people’s feeds on social media and is proposed as part of a new digital duty of care framework. The Australian Prime Minister Anthony Albanese is expected to promote this digital duty of care framework at the UN General Assembly in New York later this month, reports the Guardian.
Non-algorithmic chronological feeds: Under the proposed model, users who exercise the opt-out would see their feeds revert to chronological content only from friends, family, and groups/pages that they have chosen to follow.
Why? The move is aimed at curbing the spread of content that reinforces harmful behaviours around body image, women’s safety, and child protection. These are the areas where algorithmic amplification has been identified as a contributing factor.
Additional monitoring obligations on platforms: In addition to the algorithm opt-out function, the laws would require platforms to better identify and act on problematic or illegal content to protect vulnerable users, including children.
“I’m deeply skeptical. I fear that this is just going to be an attempt by the government to censor social media. I fear, given Labor and the minister’s past biases, that this will just be an attempt to censor the internet, and that’s not what we should be doing,” said Angus Taylor, opposition member, in an interview with The Guardian.
What comes next: The draft legislation is expected to be released as Parliament resumes this week. After placing the bill, the government may conduct public consultation and seek comments before any bill is formally introduced or enacted.
In a separate amendment introduced to Parliament, the Online Safety Amendment (Strengthening Enforcement for the Social Media Minimum Age) Bill 2026 [pdf], would double civil penalties for noncompliant platform providers from 30,000 to 60,000 penalty units. It will also expand the commissioner’s information-gathering powers. This Bill lets the commissioner go beyond the platform itself and compel third parties such as age verification vendors, contractors, and others involved in a platform’s age-assurance processes.
Why it matters: After Meta temporarily blocked PM Modi’s reel, when the company was summoned, it was also questioned for algorithmic bias too. Similarly, in February 2025, French prosecutors opened an investigation into X over allegations its algorithms distorted public discourse. Algorithmic transparency is one of the major concerns in platform regulation as it has real-world consequences, including misinformation or child harms.
For instance, before Elon Musk acquired X, a 2021 Twitter study (pdf) found right-leaning content received greater amplification across most countries examined.
In Feb 2026, Meta launched “Dear Algo,” an AI-powered feature on Threads that lets users directly influence their feed by publishing a post starting with “Dear Algo” followed by what they want to see more or less of. The platform’s AI systems interpret the request and adjust the user’s feed for three days. When it was launched, the feature had rolled out in the US, UK, Australia, and New Zealand, with plans to expand to more countries. In testing, MediaNama found the feature initially worked as intended but subsequently began surfacing unrelated posts in Tagalog for an UK-based user.
In an attempt to give algorithmic flexibility to the users, Elon Musk had announced that X’s algorithm will become “purely AI” by November or December 2025. Users will be able to adjust their feeds dynamically by prompting Grok, the platform’s AI chatbot. When MediaNama queried Grok bot regarding this feature, it cautioned.
“Results aren’t always instantaneous or perfect (the system still balances many signals, and availability of full natural-language control can vary), but asking Grok is now one of the most direct ways to steer your personal algorithm. Be specific for the best effect,” said the Grok chatbot.
Why chronological feeds are bad business for social media platforms? A Meta-backed study, published in the journal Science, ran in September-December 2020. It served reverse chronological feeds to approximately 7,200 US adults on Facebook and 8,800 on Instagram, instead of the standard algorithmically ranked feeds.
The findings from the study suggested the users on chronological feeds got bored faster and left for rival platforms. Instagram app users spent nearly eight hours less on Instagram over three months, increasing their time on TikTok by 36% and YouTube by 20%.
The findings line up with at least two earlier internal Facebook experiments from 2014 and 2018, which also found that chronological feeds drove users away. [You can read TechDirt’s report based on Facebook’s internal documents leaked by whistleblower Frances Haugen.]
The Meta-backed study also stated that switching to a chronological feed did not meaningfully change users’ political knowledge, attitudes, or behavior, including whether they attended a protest or voted.
Also Read:
