Summary

  • Australia’s 15 October exposure draft would require covered platforms to offer logged-in users a persistent choice between algorithmically selected material and a feed drawn from accounts they follow.
  • The proposed “following-only” feed is not ad-free: the draft expressly permits advertising, including from accounts the user does not follow.
  • A privacy notice for the feed-rule consultation says AI may support a separate Digital Duty of Care Bill consultation. The page does not clarify whether that statement also covers feed-rule submissions; it is a notice-scope question, not evidence that those submissions are processed by AI.

A choice that must remain a choice

On 15 October, the Department of Infrastructure, Transport, Regional Development, Communications, Sport and the Arts opened consultation on the exposure draft of the Online Safety (User Empowerment Tools) Determination 2026. The consultation runs until 11:59 pm AEDT on 5 November. The instrument is marked “DRAFT ONLY—NOT FOR SIGNATURE,” and the department says the enabling Online Safety Amendment (Digital Duty of Care) Bill 2026 must receive Royal Assent before the Minister can determine the rule. The proposal is therefore not a duty that platforms must meet today.

The draft sets out more than a pair of buttons. A covered service would have to make the tool obvious, accessible, prominent and functional, keep it available, explain it in clear language and avoid making the service harder to use or discouraging a selection. The setting could not be changed during or between uses without the user’s explicit choice. If no preference had been recorded when the rule commenced, the draft would prompt the user on their next logged-in visit.

For covered social-media services and age-restricted platforms that provide user-generated content through a recommender feature, the proposed choice is between algorithmically selected material based on user profiling and content from accounts the user has chosen to follow. That distinction gives users a direct control over one important source of feed material. It does not, on the text available, promise a feed free from every other commercial or ranking mechanism.

“Following-only” still has an advertising lane

The most consequential qualification is explicit. The draft allows advertising in a following-only feed, including advertising from an account the user does not follow. So “following-only” describes the ordinary content-source rule, not an ad-free environment. The provision preserves a route for material outside a user’s follow graph even after that user rejects the algorithmically selected feed option.

The draft does not, in the cited provisions, explain how advertising would be selected, ranked or personalized in that feed. That omission should not be filled with assumptions about current platform systems. It does, however, make a useful consultation question: what information and controls will let a person distinguish followed-account posts from paid material, and will the final instrument say anything about how those ads are delivered?

The submission channel needs a clearer boundary

The consultation page says submissions will inform development of the user-empowerment rule. It asks contributors to provide contact details and choose whether a submission may be published or kept confidential; public submissions may appear in part or full, with personal contact details redacted. The page’s privacy collection notice is headed with the feed-rule instrument and explains those uses.

Its final AI paragraph changes documents. It says the Department may use technology, including AI tools, to support consultation on the exposure draft of the separate Digital Duty of Care Bill, including identifying abusive or offensive language and summarizing feedback that may contain personal information. It says those activities may generate or infer new personal information and that outputs will be reviewed by staff. The page does not say whether that AI statement also applies to submissions on the feed-rule instrument.

The public record supports identifying a scope mismatch, not claiming that these submissions are being AI-screened or summarized.

A persistent interface choice and a consultation response are different forms of user agency. The first can change what appears in one person’s feed; the second enters an institutional process that may publish, summarize or use submissions to shape a rule. Clear labeling of both the feed’s ad exception and the consultation’s data-processing scope would make those boundaries easier to understand before a person chooses or submits.

Sources