Duty of Care; Duty to Whom? The US–Australia Clash Over Online Safety
"Think of the children" is one of the most powerful phrases in politics. Sometimes it is entirely justified. Children can be bullied, sexually exploited, exposed to pornography, encouraged towards eating disorders and targeted by technological systems they scarcely understand. No serious defender of free speech needs to deny any of that.
But "think of the children" also has another political advantage. Once a proposal is presented as necessary to protect children, anyone questioning the powers being created can be made to sound as though he is defending the harm rather than questioning the remedy.
That is why the increasingly sharp dispute between the United States and Australia over Labor's proposed Digital Duty of Care deserves more attention than another round of arguments about whether social-media companies are nice people. They plainly are not charities. The real question is much older: who gets to decide what people may see, and how much power should government possess to make that decision indirectly?
The Trump administration has now made an unusually public intervention in Australia's domestic debate. Washington says it has serious concerns about the proposed duty of care and warns that it could facilitate censorship, including what it describes as the extraterritorial censorship of speech protected in the United States.
Australia's government presents the proposal as safety regulation. Communications Minister Anika Wells has stressed protecting children and shifting responsibility onto technology companies, while Anthony Albanese insists that the plan is "not about government control" but about returning control to users. One headline proposal, "My Feed, My Way," would allow Australians to choose whether they receive algorithmically recommended material or principally material from accounts they follow.
Taken by itself, giving adults more control over their feeds sounds like an expansion of freedom rather than a restriction upon it. I have no particular objection to that principle. If Facebook, TikTok or any other platform wants to construct a psychological profile of me and use an algorithm to decide what will keep me staring at the screen, giving me an effective OFF switch sounds perfectly reasonable.
But that is not the end of the bill. The wider proposal imposes a duty upon digital platforms proactively to address foreseeable harms. For children, the categories under discussion include pornography, bullying, eating-disorder material, misogynistic content and glorification of crime. For adults, the government's stated categories are much narrower and largely connected with existing illegality. Yet the draft would also give the communications minister power to add categories of harm by regulation, provided the statutory requirements were met, with such regulations capable of parliamentary disallowance.
That is where the American objection becomes important. What is "harm"? Who defines it? Who decides whether a platform has done enough to prevent it? And what happens when the safest commercial response to an ambiguous regulatory obligation is simply to make controversial material harder to find?
Censorship in the digital age does not necessarily require a policeman arriving at the newspaper office and confiscating the printing press. A platform can leave a post technically online while ensuring almost nobody encounters it. It can demonetise it, remove it from recommendations, suppress it in searches or prevent an algorithm from distributing it. The speaker has not formally been silenced. He has merely been placed in a room where nobody can hear him.
This is one of Washington's more interesting objections. If governments impose sufficiently broad duties upon platforms to prevent vaguely defined harms, rational companies will protect themselves. Faced with enormous potential penalties, their incentive will not be to conduct seminars on John Stuart Mill. It will be to remove, suppress or demote borderline material.
The censorship, in other words, need not be explicitly ordered by government. Government establishes the risk. The corporation performs the suppression. That distinction may be legally important, but it will not be terribly comforting to the person whose lawful opinion has disappeared from everybody's feed.
This is why the child-safety framing deserves particular scrutiny. There are genuine and horrifying online harms to children. Australia already has an extensive eSafety apparatus dealing with child cyberbullying, image-based abuse, illegal material and other serious online harms. Nobody needs to invent those problems; they exist.
But legislation justified by extreme cases has a habit of surviving the extreme case. Powers created to deal with child pornography do not necessarily remain confined to child pornography. Concepts developed to deal with suicide promotion, sexual exploitation and serious bullying can gradually expand into broader ideas of psychological harm, social harm and harmful information. The crucial question is therefore not whether today's minister promises sensible administration. It is what powers today's Parliament leaves for tomorrow's minister.
That is a basic free-speech principle: never judge a coercive power solely by imagining it in the hands of politicians you trust. Imagine it in the hands of politicians you distrust.
Australia and the United States approach this problem from markedly different constitutional traditions. The United States begins with a powerful presumption against government interference with speech. Australia has no equivalent constitutional guarantee of free speech. Australians therefore have good reason to pay attention when Americans look at one of our regulatory schemes and see possibilities that our own political culture has become accustomed to overlooking.
That does not mean Washington is acting from pure constitutional principle. American technology companies have billions of dollars at stake. The Trump administration openly objects to foreign taxes, fines and regulatory burdens directed at US technology companies, and Washington has asked that American social-media companies be excluded from parts of Australia's proposed regime.
Australia should reject that exemption. If a law is legitimate, it should not mysteriously cease to be legitimate when the corporation affected happens to be American. And if it is dangerous, Australians deserve protection from it too. The US commercial argument and the US free-speech argument therefore need to be separated. One may be self-interested while the other remains correct.
The Australian government's defence also contains a legitimate point. Social-media algorithms themselves exercise enormous power over speech. Facebook or TikTok deciding what millions of Australians see is not some pristine state of liberty untouched by human interference. Private corporations already amplify some voices and bury others, often through systems users barely understand. Requiring genuine user choice over algorithmic feeds can therefore be defended as increasing individual autonomy.
Fine. Give users that choice. But user choice and government-defined harm are different questions. A genuine "My Feed, My Way" principle would place the citizen in control. Let me decide whether I want algorithmic recommendations. Let parents employ strong filters for their children. Give users straightforward tools for blocking pornography, violence, gambling material or whatever else they do not wish to encounter.
The freedom principle should be simple: maximise the individual's ability to choose what enters his screen while minimising the government's ability to decide what may enter everybody's screen.
That is where the Australian proposal needs to be tested. The government naturally reaches for children because the strongest examples favour regulation. Show parents a suicidal teenager being repeatedly fed self-harm material and few will respond with an abstract lecture about the marketplace of ideas. Nor should they. But legislation must be judged at its boundaries, not merely by its most sympathetic example.
The test case for free speech is never the material everybody agrees should be stopped. It is lawful material that officials consider harmful, offensive, irresponsible or socially undesirable. If the architecture being constructed can reach that material, then we are no longer talking simply about protecting children from pornography and bullying. We are talking about regulating the boundaries of permissible discourse.
This is why the American intervention is useful even for Australians who have no desire to take political instructions from Washington. It forces the debate onto terrain Canberra would understandably prefer to avoid.
Not: do you care about children? Of course, we do. Not: are technology companies capable of causing harm? Of course, they are. The harder questions are these: how precisely is harm defined, how far does the definition extend beyond existing illegality, how easily can those categories be expanded, what protections exist for lawful controversial speech, and who bears the burden when government pressure causes a private platform to suppress something nobody has actually made illegal?
Those questions cannot be answered with another photograph of a worried parent. Nor can they be answered by saying that the present government has good intentions; given its track record, that would be foolish. Most restrictions upon liberty arrive accompanied by good intentions. Governments rarely announce that they are seeking unnecessary powers to be used irresponsibly at some future date.
The Australian instinct increasingly seems to be: identify a social problem, create a regulator and give it sufficient powers to make the problem go away. The older American instinct is more suspicious: what happens when the regulator itself becomes the problem? On free speech, a little of that American suspicion would do Australia no harm.
Protect children, certainly. Prosecute genuine criminality. Give parents better tools. Give adults control over their algorithms. Force technology companies to be transparent about what their systems are doing. But when government moves from protecting people against clearly defined unlawful conduct towards deciding what lawful information is sufficiently "harmful" to require suppression, friction or disappearance, the burden of proof should change sides.
Freedom does not require trusting Big Tech. It requires distrusting concentrated power wherever it resides. That includes Silicon Valley. And it includes Canberra.
https://www.theepochtimes.com/world/us-warns-australias-digital-duty-of-care-law-risks-viewpoint-based-censorship-6093674
