US Embassy Calls Australia's Algorithm Opt-Out Proposal a Censorship Risk

Key Takeaways

• The US embassy in Canberra submitted formal objections to Australia's draft Digital Duty of Care laws, saying the proposal to let users opt out of algorithms amounts to "censorship of protected speech" and risks "viewpoint-based censorship" through vague harm definitions.
• The US warned that mandated platform design features applied to algorithms could affect what users see globally — not just in Australia — and that pre-emptive content demotion by platforms could reduce the reach of independent journalists and minority viewpoints.
• Australian Prime Minister Anthony Albanese, speaking at the UN General Assembly during his visit to the US, responded that the legislation is about giving individuals control over what they receive on their devices, not giving government control over content.
• The Digital Duty of Care bill — currently at exposure draft stage with a feedback deadline of September 22nd — bundles an algorithm opt-out provision with a broader duty of care enforced by the eSafety Commissioner, carrying fines exceeding $100 million.
• Australia introduced a social media ban for under-16s last December; this bill is the next step in its online safety legislative program.

What the US Is Objecting To

The US embassy's submission to Australia's draft Digital Duty of Care laws is a rare move — formal diplomatic objection to proposed domestic legislation from a close ally. The objections, relayed by the BBC's Sydney bureau, center on two connected concerns.

The first is the definition of harm. The bill would allow the eSafety Commissioner to fine platforms more than $100 million if they fail to meet a duty of care standard, with the definition of what constitutes insufficient moderation left largely to the Commissioner's discretion. 

The US submission asked Australia to clarify how harm and risk will be determined, warning that enforcing vague definitions could lead to viewpoint-based censorship — decisions shaped by the preferences of whichever official happens to hold the role rather than by clear legal standards.

The second concern is the algorithm opt-out provision itself. The US argued that mandated platform design features, particularly around algorithms, could affect users globally (not just in Australia) if tech firms have to build compliance into systems that operate across borders. 

It also raised the possibility that platforms, facing legal liability for content that reaches users through their recommendation systems, would pre-emptively demote independent journalists or minority viewpoints to avoid being found in breach. An opt-out option, the submission argued, could allow regulators to impose rigid, one-size-fits-all design requirements on platforms operating in very different contexts.

Albanese's Response and the Tension It Reveals

Australian Prime Minister Anthony Albanese, speaking to the media in New York during his visit to the UN General Assembly, pushed back directly. The legislation, he said, isn’t about giving the government control over content — it’s about giving individuals control over what they receive on their devices. 

That framing positions the algorithm opt-out as a user empowerment measure rather than a regulatory intervention, and it’s not wrong on its own terms: the ability to choose a chronological feed over an engagement-optimized one is a genuine user right that most platforms have resisted implementing prominently even when they offer it.

The tension between those two framings — user empowerment versus regulatory overreach — is real, and both sides are identifying genuine risks. The US is right that vague harm definitions enforced by a single regulator with nine-figure fining power create conditions for politically shaped content decisions. Albanese is right that algorithmic recommendation systems that optimize for engagement over accuracy or wellbeing are causing documented harm, and that users should be able to choose otherwise.

Where it gets complicated is in the bill's design. As we covered when the exposure draft was published on September 8th, the algorithm opt-out is bundled with a duty of care provision that does not define what constitutes insufficient moderation with any precision. 

The opt-out provision is genuinely popular and largely benign. The duty of care provision is where the authority to shape content lives. The US submission does not fully distinguish between the two, which weakens its censorship argument. But the argument isn’t wrong — it’s aimed at the right half of the bill.

What This Exchange Actually Reveals

The US position isn’t straightforward to assess, because it’s simultaneously a principled free speech argument and a statement of commercial interest. The platforms most affected by Australia's proposed laws — Meta, Google, TikTok — are US-headquartered companies. 

The Trump administration has been consistently skeptical of foreign regulatory frameworks that impose design obligations on US tech companies, and the submission's reference to Australia's "reputation as a jurisdiction that enables innovation" reads as diplomatic language for "this is bad for US tech business." That doesn’t make the legal argument wrong. It does make the motivation worth noting.

What the exchange illustrates most clearly is a dynamic we have written about in the context of the Australian bill: when a law bundles a popular, largely symbolic reform with a vague enforcement power, the vague enforcement power is the story. 

The US submission is aimed at the duty of care provision's potential for misuse. Australia's government is defending the algorithm opt-out. Both are talking about the same bill, and neither is quite engaging with the other's actual concern. That’s a sign the bill's design is creating more confusion than it resolves — which is exactly the argument for separating the two provisions before it passes.