Opinion: If a government wants more power over the digital information environment, there may be no more politically effective way to ask for it than to begin with children.

Protect children from pornography. Stop grooming. Reduce cyberbullying. Prevent algorithms feeding vulnerable teenagers eating-disorder material, self-harm instructions or violent content.

Who seriously wants to campaign against those objectives?

That is why Labor’s new Digital Duty of Care deserves to be examined not only by asking whether its intentions sound admirable, but by reading the powers contained in the exposure draft itself.

And once that is done, the debate becomes considerably more complicated.

Anthony Albanese says the reform is not about government control.

Communications Minister Anika Wells says it is about safety standards and forcing enormously profitable technology companies to accept responsibility for products capable of causing real harm.

On much of that, they have a persuasive case.

Big Tech should not enjoy an automatic exemption from ordinary expectations of product safety simply because its products appear on screens rather than supermarket shelves.

But Australians should be equally wary of another proposition: that because some of the harms being targeted are unquestionably terrible, every power attached to the solution must therefore be benign.

That does not follow.

The easy part of the legislation is easy to support

Labor’s exposure draft begins with categories few reasonable people would defend.

For adults, its definition of seriously harmful material includes child sexual exploitation and abuse, grooming, encouragement of sexual or extreme violence, explicit rape and death threats, serious menacing or harassment, suicide and self-harm encouragement, terrorism-related material, instruction in criminal offending and promotion of illicit drug use.

For children, additional categories include pornography, promotion of disordered eating, bullying, dangerous stunts, glorification of crime and material promoting hostile attitudes towards women or gender equality.

Those are not descriptions of somebody criticising Anthony Albanese’s tax policy.

They are genuine categories of online harm.

The draft also says lawful private communications solely between consenting adults are not required to be acted upon under the duty.

This is why claims that the bill, exactly as written today, simply bans ordinary criticism of Labor are too broad.

It does not.

The stronger concern lies elsewhere.

The list Parliament sees today does not have to remain the list tomorrow

Section 25C is where the political sales pitch meets the institutional question.

After listing the serious harms currently proposed for adults, the draft adds another category.

It permits the Communications Minister, through a legislative instrument, to determine additional material or conduct that the minister is satisfied may cause serious harm.

The child-harm provisions contain a comparable power.

The draft also allows the minister to designate additional design features as having negative behavioural impacts.

This matters because Parliament would not necessarily need to pass a fresh amendment to the primary Act every time the regulatory perimeter moved.

A minister could use delegated legislation.

There are safeguards.

The minister must seek advice from the eSafety Commissioner before exercising several of these rule-making powers and must have regard to that advice.

Wells says the resulting instruments are disallowable, meaning Parliament can overturn them.

That is meaningful oversight.

But it is not the same thing as requiring Parliament to positively pass a new bill before a new category of expression enters the regulatory framework.

That distinction is exactly why Human Rights Commissioner Lorraine Finlay has warned about “open-ended ministerial discretion”.

She supports the concept of a duty of care while arguing powers affecting fundamental rights need clear boundaries, transparency, parliamentary scrutiny and review.

That is hardly an extremist objection.

It is the kind of objection Australians should want a Human Rights Commissioner to raise before legislation is finalised.

Today’s minister is not the only minister who matters

Whenever governments defend discretionary powers, the discussion quickly becomes personal.

Trust us.

We would never use it that way.

That is the wrong test.

The question is not whether Anika Wells intends to declare criticism of Labor a serious harm next Tuesday.

There is no evidence she does.

The question is what powers Australian law should place permanently in the hands of whoever occupies the office in the future.

One day that minister will be Labor.

Another day it may be Liberal, National, One Nation or somebody from a political movement that does not yet exist.

A sound civil-liberties test is therefore simple: would supporters of this legislation be equally comfortable giving the same discretion to the politician they trust least?

If the answer changes with the name on the ministerial door, the safeguard is probably not strong enough.

Labor says adults keep their freedom

Albanese has been explicit.

Asked whether the laws could allow eSafety or government regulators to police misinformation or political commentary, he said they were not about censorship.

Wells said the government could emphatically rule that out.

She has even used the example that Australians remain free to believe the moon landing was fake.

Those assurances matter.

So does the current wording of the listed adult harms, which does not simply nominate “misinformation”, “offensive politics” or “criticism of government” as prohibited categories.

But governments legislate through words, not press conference intentions.

If Labor genuinely has no interest in allowing ordinary political communication to be captured, Parliament should be able to write that protection as strongly as possible into the final statute.

Clear protections for political communication would strengthen the government’s case rather than undermine it.

‘My Feed, My Way’ is not the strongest censorship argument

Critics should also distinguish between different parts of Labor’s package.

The “My Feed, My Way” proposal is difficult to attack as an obvious form of government censorship.

Under Labor’s plan, social media users aged 16 and over would be offered a choice about their primary feed.

They could retain algorithmically recommended material or select a feed focused on friends and creators they deliberately follow.

They could change their preference later.

That is, at least in principle, an expansion of user choice.

For years technology companies have decided which posts receive priority through opaque recommendation engines designed around engagement.

Giving users a meaningful escape from those systems can reasonably be described as returning control to individuals.

There is no contradiction in supporting that feature while opposing excessively broad ministerial discretion elsewhere in the same bill.

Legislation does not need to be either completely good or completely sinister.

A serious Parliament should be capable of keeping the useful parts and tightening the dangerous ones.

The deeper change is the duty imposed on platforms

The proposal goes far beyond a feed toggle.

Online services would have to ensure, so far as reasonably practicable, a safe online environment.

They would have mandatory obligations to manage design features appropriately, conduct written assessments of foreseeable risks and take effective measures to address the risks identified.

That means government regulation would reach into the architecture of digital products.

Recommendation systems.

Endless feeds.

Feedback features.

Time-limited features.

Other designs capable of influencing behaviour.

There is a legitimate argument for doing exactly this.

An online platform can cause harm through design even when no single post is obviously illegal.

A recommender system can repeatedly push vulnerable users towards more extreme material because that content keeps them engaged.

An endless feed can be engineered to maximise time spent on an app.

Child-safety advocates are right to say regulating individual posts alone can become a hopeless game of whack-a-mole.

But when government moves from policing specific illegal content towards imposing broad systemic duties, the quality of definitions and oversight becomes even more important.

The financial pressure on platforms will be enormous

Failure to comply with the Digital Duty of Care carries a civil penalty of 60,000 Commonwealth penalty units.

With the current penalty-unit value and the rules applying to corporations, the maximum exposure for a large company can reach $109.2 million.

The government has also said overlapping breaches involving other online-safety requirements could take penalties above $200 million in some circumstances.

Labor sees those numbers as necessary because companies such as Meta, Google and other global platforms cannot be meaningfully deterred with small fines.

That is understandable.

But large penalties can also create another risk.

If the boundary of prohibited or regulated harm becomes uncertain, corporations will have an incentive to remove too much rather than too little.

A platform facing a nine-figure penalty may not wait for a court to establish the most speech-protective interpretation of an ambiguous rule.

It may simply suppress the material.

The government therefore does not need to personally censor a political statement for regulation to produce a chilling effect.

The design of incentives can do some of the work.

That does not mean such an effect is inevitable.

It means precision in the final law matters enormously.

The Coalition’s language is overheated, but its question is legitimate

Opposition Leader Angus Taylor has now declared the exposure draft unacceptable in its current form and called it a “blank cheque for political censorship”.

That phrase runs ahead of what the bill currently does.

No blank cheque is literally being handed over.

Ministerial instruments face legal requirements, eSafety advice and parliamentary disallowance.

Nor does the present draft explicitly outlaw ordinary political criticism.

But dismissing the Coalition merely because its rhetoric is dramatic would be another mistake.

The question beneath the slogan is valid.

Why should an elected minister have an open-ended ability to add new categories of serious harm when the subject matter can affect lawful expression?

Could the criteria be narrower?

Could additions require a higher evidentiary threshold?

Could independent review be mandatory?

Could certain categories of lawful political communication receive stronger statutory protection?

Could Parliament be required to affirm particularly significant expansions rather than merely having the option to disallow them?

Those are constructive questions.

There is still time to fix the bill

The most important thing to remember is that Australians are not yet living under this Digital Duty of Care.

The document released on September 8 is an exposure draft.

Consultation remains open until noon on September 22.

The government can change it.

Parliament can change it.

That is exactly what exposure drafts are supposed to be for.

Labor should use that opportunity rather than treating every civil-liberties objection as evidence that critics want children exposed to online harm.

They do not have to choose.

Australia can require technology companies to design safer products.

It can make child sexual abuse, grooming and dangerous algorithmic amplification harder to distribute.

It can give users meaningful control over recommendation feeds.

And it can simultaneously insist that lawful political speech sits behind a high and clearly defined barrier before ministers or regulators can interfere with it.

Child protection should be the reason for better drafting, not weaker scrutiny

The political genius of attaching broad regulation to child safety is that opponents can easily be made to look as though they oppose safety itself.

That would be a dishonest way to conduct this debate.

There are parents who have lost children and advocates who have seen the devastating consequences of cyberbullying, self-harm material and predatory online systems.

Their case for action deserves respect.

But legislation written in response to heartbreaking cases still has to survive the ordinary tests applied to government power.

Who decides?

According to what definition?

What happens when definitions change?

What independent checks exist?

What protection is available for lawful speech?

What happens when a future government uses the same powers with very different intentions?

Those questions are not an attack on online safety.

They are part of designing it responsibly.

Labor is correct that Big Tech already exercises immense power over what Australians see.

Opaque algorithms can shape attention, amplify outrage and decide which voices become visible.

That private power should not be romanticised as freedom.

But reducing the power of technology companies does not require Australians to become casual about expanding the power of government.

The objective should be to give individuals more genuine control while keeping both corporate and political power constrained.

That is the standard against which the final Digital Duty of Care should be judged.

If Labor wants Australians to believe this is truly a duty of care rather than a future duty of control, the best response is not another reassurance from a podium.

It is a final law written tightly enough that Australians do not have to rely on reassurance at all.