Prime Minister Anthony Albanese has rejected claims Labor’s proposed Digital Duty of Care could be used to silence political criticism online, insisting the legislation is designed to make technology companies responsible for safety while giving Australians greater control over what appears in their social media feeds.

But the reassurance has not ended an escalating free-speech fight in Canberra.

By Thursday, the Coalition had formally declared the exposure draft unacceptable in its current form, arguing that proposed ministerial powers to expand the categories of material captured by the regime could eventually be used to regulate lawful political speech.

The dispute has created a more complicated picture than either side’s political slogans suggest.

The draft does not presently identify criticism of the government, ordinary political commentary or misinformation as prohibited adult political speech.

Its core obligations instead focus on forcing digital services to assess systemic risks, address seriously harmful material and redesign features capable of exposing users — particularly children — to foreseeable harm.

At the same time, the proposed law does contain rule-making powers allowing the Communications Minister to expand parts of the framework through legislative instruments.

Those instruments would face parliamentary scrutiny and could be disallowed, but critics argue the initial discretion is still too broad.

Albanese directly rejects censorship claim

The Prime Minister was challenged on the issue when he and Communications Minister Anika Wells unveiled the draft legislation at Parliament House on September 8.

A reporter asked whether the legislation could allow the eSafety Commissioner or another regulator to police misinformation, political statements or commentary.

Albanese responded that the proposal was not about censorship but about allowing individuals to control their own online experience.

Wells went further, saying the government could emphatically rule out the suggestion raised in the question.

Labor’s central argument is that the law targets the systems through which online harm develops rather than establishing a government body to decide which political opinions Australians are permitted to hold.

The government describes that as a shift from reactive regulation to prevention.

Under the current model, regulators frequently intervene after harmful material has appeared and somebody has complained.

A Digital Duty of Care would instead require services to assess foreseeable risks in advance, document those risks and take reasonable steps to reduce them.

That approach places more responsibility on companies including social media networks, messaging platforms, online games, search engines, app stores and certain artificial intelligence services.

What the draft actually defines as serious harm

The September exposure draft provides a detailed initial list of material and conduct regarded as seriously harmful.

It includes child sexual exploitation and abuse material, grooming of children for sexual activity, encouragement of sexual violence or extreme violence, explicit death or rape threats and conduct intended to seriously harm somebody through menacing or harassment.

It also includes material encouraging suicide or self-harm, support for listed terrorist organisations, encouragement of terrorist acts, instruction in criminal offences, promotion of illicit drug use and certain abhorrent violent conduct.

Those categories are substantially different from ordinary arguments over tax, immigration, climate policy, political parties or the performance of the Albanese government.

There is no clause in the present list simply declaring criticism of politicians, dissent from government policy or politically embarrassing information to be serious harm.

That is the strongest basis for Albanese’s rejection of claims that Labor is openly legislating a political censorship regime.

However, it is not the end of the free-speech question.

The ministerial power at the centre of the controversy

Section 25C of the exposure draft would allow the Communications Minister, through a legislative instrument, to determine additional material or conduct that the minister is satisfied may cause serious harm.

A similar mechanism exists for material harmful to children.

The minister would be required to seek advice from the eSafety Commissioner before making such an instrument and have regard to that advice.

The resulting rules would also be disallowable instruments, meaning Parliament could overturn them.

Government material says the rules must be considered alongside freedom of expression and Australia’s human-rights obligations.

Wells argues those safeguards make descriptions of an unchecked censorship power misleading.

She has also noted that delegated rule-making powers already exist elsewhere in the Online Safety Act and are not a completely new regulatory invention.

But the Coalition says the problem is what a future minister might try to add.

Opposition Leader Angus Taylor said on Thursday the exposure draft gave the government what he described as a “blank cheque for political censorship”.

Shadow communications minister Sarah Henderson argued there were insufficient safeguards around the discretion.

The Human Rights Commissioner, Lorraine Finlay, has separately raised concerns about open-ended ministerial powers and called for clear boundaries, transparency and meaningful parliamentary oversight.

Those concerns do not prove Labor intends to censor political opponents.

They do establish that the argument is about more than a misunderstanding of the policy.

The question is whether a law drafted for genuine and serious online harms contains sufficient protections against future expansion.

Adults and children are treated differently

Another source of confusion is that the draft does not use exactly the same harm threshold for adults and children.

Wells says the adult part of the duty is focused on seriously harmful material and conduct, much of it already linked to illegal behaviour or severe threats.

Children receive broader protection.

The exposure draft specifically identifies pornography, promotion or instruction in disordered eating, material encouraging hostile attitudes towards women or gender equality, glorification of crime or dangerous stunts, and abuse, harassment or bullying as potentially harmful to children.

The government argues that distinction is necessary because minors are more vulnerable to design features and content that might not justify the same regulatory response for an adult.

Digital platforms would also have to examine how their own product design contributes to those risks.

The legislation specifically identifies recommender systems, endless feeds, engagement feedback and time-limited features as areas capable of affecting behaviour.

This is significant because the law is not designed solely around deleting individual posts.

It reaches into the architecture used by platforms to keep people watching, clicking and scrolling.

‘My Feed, My Way’ attacks the algorithm problem from another direction

The most consumer-facing part of Labor’s plan is branded “My Feed, My Way”.

Australians aged 16 and over would be given an ongoing choice about the type of default feed they want to see.

One option would permit personalised recommendations generated by the platform’s algorithm.

The other would allow users to move away from personally recommended content and concentrate on friends, creators and accounts they have deliberately chosen to follow.

The government says platforms would need to respect that selection rather than repeatedly steering people back towards algorithmically chosen recommendations.

The proposal reflects growing concern about recommendation engines because users do not simply search for every post they encounter.

Platforms decide which material to push towards them.

A person who engages with one type of content can quickly be shown more of the same, including increasingly provocative material if the algorithm predicts it will hold attention.

Labor argues individual Australians should be able to weaken that influence without abandoning the platform entirely.

It is therefore difficult to characterise the algorithm-choice provision itself as government control over political speech.

Its stated mechanism gives the user more control, not a minister the ability to choose the individual posts appearing in that user’s feed.

Platforms would face a legal duty, not liability for every bad post

The draft also makes an important distinction between systemic failures and isolated pieces of content.

A platform would not automatically breach the duty simply because a harmful post appeared.

Instead, providers would be required to maintain effective systems and processes, conduct risk assessments and take reasonable measures against foreseeable risks.

Repeated harmful content, ineffective mitigation or a systemic failure to address known risks could become evidence that the service had breached its duty.

That model resembles safety regulation in other industries.

A company is expected to identify risks inherent in the product it operates and take reasonable precautions rather than merely responding after somebody is injured.

Wells has repeatedly compared the concept with safety standards applied to products such as cars, food and toys.

The penalties give the debate real stakes

This is not intended to be a voluntary code.

The government says failure to meet the Digital Duty of Care could attract penalties of up to $109.2 million.

The eSafety Commissioner would be responsible for compliance and enforcement.

That size of penalty is part of the reason the wording matters so much.

If a platform faces a substantial financial risk, critics argue it may remove or suppress more material than legally necessary rather than test the boundary of an ambiguous harm category.

This phenomenon is sometimes described as over-removal or a chilling effect.

The government’s counterargument is that freedom-of-expression considerations, the serious-harm threshold, regulatory oversight and parliamentary disallowance provide safeguards against that outcome.

The dispute therefore is not merely about whether a minister personally deletes a Facebook post.

It is also about how companies will behave when broad legal duties are combined with very large penalties.

Coalition hardens its position

The political situation shifted further on September 10 when Taylor confirmed the Coalition would not support the exposure draft in its existing form.

The opposition says it supports protecting children and imposing greater accountability on major technology companies.

Its objection is that those goals should not be used to justify open-ended regulatory discretion over lawful speech.

That puts Labor in the unusual position of defending a policy whose central concept — making platforms more responsible for online safety — has substantial cross-party sympathy, while fighting over the exact scope of government power used to achieve it.

The Coalition’s stance also comes while Parliament is debating separate measures strengthening enforcement of Australia’s under-16 social media regime.

The two policy packages have sometimes been discussed together, but they are not identical.

The Digital Duty of Care released on September 8 remains an exposure draft undergoing consultation.

It has not yet completed the parliamentary process and has not become operative law.

The draft can still change

This is perhaps the most important fact for Australians trying to judge the controversy.

The government is seeking feedback on the exposure draft until noon on September 22.

Labor says final legislation will be introduced to Parliament later this year.

That creates an opportunity for the ministerial powers, definitions of harm and safeguards to be narrowed, clarified or otherwise amended before MPs vote on the final bill.

It also means some political claims about what the legislation will inevitably do are premature.

The relevant text is still being negotiated.

There is a credible case for stronger online safety regulation.

There is also a credible case for demanding unusually precise safeguards whenever government creates powers capable of affecting lawful expression.

Those positions are not mutually exclusive.

Australians can support action against child exploitation, grooming, death threats and algorithms pushing vulnerable children towards self-harm while still asking Parliament to ensure a future minister cannot stretch the concept of harm into ordinary political disagreement.

Albanese’s censorship denial can be tested against the final law

For now, the Prime Minister’s statement is clear.

He says the reform is not designed to give his government control over what Australians say.

Its stated purpose is to shift responsibility towards the technology companies designing online environments and give individuals more power over what those systems recommend to them.

The current exposure draft provides substantial support for that description.

Its listed adult harms concern extremely serious material and conduct, not ordinary political criticism.

Its algorithm proposal gives users a choice rather than prescribing a government-selected feed.

But the Coalition is also correct about one narrower factual point: the draft gives the Communications Minister a mechanism to add categories of harm in the future.

That power comes with eSafety advice and parliamentary disallowance, but Parliament will now have to decide whether those safeguards are strong enough.

The real test of Labor’s promise therefore will not be the Prime Minister’s press conference.

It will be the final text that eventually reaches Parliament.

If the legislation remains tightly directed towards objectively serious harms and systemic platform safety, Albanese’s claim that the reform is about protection and user control will be easier to sustain.

If future wording leaves political speech vulnerable to broad discretionary definitions, the opposition’s warnings will become harder to dismiss.

For Australians, that is the distinction worth watching as one of the country’s most ambitious attempts to regulate Big Tech moves from political announcement to actual law.