Elon Musk Was Right About eSafety

Elon Musk was right about Australia’s under-16 social media ban.

When the Albanese Government first announced it, Musk called the scheme a “backdoor way to control access to the internet by all Australians”. It was easy for the government to dismiss that as another billionaire protecting his own patch.

Less than two years later, Labor wants to give Julie Inman Grant’s eSafety office the power to demand board minutes, internal emails, company directions, and other confidential documents while enforcing the ban. The demands would extend beyond social media companies to age-assurance providers, app stores, contractors, and potentially ordinary people who hold information eSafety considers relevant.

The wording of the proposed law matters. Section 63G would allow the Commissioner to issue a notice to “a person”, rather than limiting the power to the company under investigation. The Institute of Public Affairs and X have both warned that this could catch users, employees, contractors, and other third parties.

Refusing a demand would be costly. The bill doubles the maximum civil penalty from 500 to 1,000 penalty units. Since a Commonwealth penalty unit increased to $364 on 1 July, an individual could be fined up to $364,000. For a company, the maximum could reach $1.82 million. For systemic breaches of the under-16 ban, the corporate maximum under the bill would actually be $109.2 million.

Those are court-imposed maximums, of course, rather than automatic fines. Even so, threatening a company with more than $100 million, or an individual with hundreds of thousands, is hardly a polite request for assistance.

Labor argues that the companies have been taking the mickey. Communications Minister Anika Wells says they are frustrating the ban, doing the bare minimum, and forcing eSafety to accept their claims without seeing the records behind them.

Now, as a father, I have no desire to see children hooked on deliberately addictive platforms, bullied online, or fed sexual and violent rubbish by an algorithm. The big social media companies have spent years pretending that none of this is really their responsibility. The trouble is that Labor’s ban has not fixed it.

The government says more than five million accounts linked to under-16s have been removed, restricted, or deactivated since the law took effect on 10 December 2025. Five million sounds impressive, but these are accounts, not children. A single child might have accounts on Facebook, Instagram, TikTok, Snapchat, and YouTube.

And eSafety’s own research gives a better idea of what has happened. Roughly seven in ten parents whose children had accounts on Facebook, Instagram, Snapchat, or TikTok before the ban said their children still had them afterward.

The most common reason was fairly mundane. In 66.8 percent of cases, the platform had never asked the child to verify their age. A second report published by eSafety found no significant decline in how often children aged 10 to 15 reported using social media. They were still spending an average of two hours and nine minutes on it during a school day, rising to almost four hours on weekends.

Worse, parents had less idea what their children were doing. Before the ban, 23.3 percent of parents were unaware that their child had recently used social media. Three months later, that figure had risen to 33.3 percent.

Among parents of children aged 10 to 12, the increase was even sharper. The proportion who did not know their child was using social media rose from 35.4 percent to 49.1 percent.

The report does not prove that the ban caused children to hide their social media use, but it raises the rather obvious possibility. Teenagers have always found ways around rules, and a government announcement does not suddenly change that. The ban may have pushed some of this activity further from parental view while changing very little about how much time children spend online.

If this were almost any other policy, figures like these would prompt a rethink. In Canberra, failure usually leads to demands for more money and greater powers.

X has told the Senate inquiry that the new powers do not pay enough regard to procedural fairness, privacy, the effect on online services, or Australia’s digital economy. Its strongest objection concerns people who are not even in Australia.

According to X’s submission, the bill could allow eSafety to compel someone overseas to hand over information merely because that person is “affiliated” with a company operating here. X says this conflicts with international legal principles and could have a “severe impact on international comity”.

International comity is not a phrase likely to come up over a beer, but the idea is straightforward. Countries are expected to show some respect for one another’s laws, courts, and territorial limits.

We would be furious if Beijing demanded emails from an Australian employee because his company had Chinese customers. The same would be true if a European or Middle Eastern regulator ordered an Australian citizen to surrender private records over material that was perfectly lawful here. Canberra would (hopefully) tell them to get stuffed, and rightly so.

Our government cannot claim for itself a power that it would never accept in the hands of another country.

X is protecting its commercial interests. But having a financial interest in a dispute does not automatically make an argument wrong, and Julie Inman Grant has already shown that she thinks an Australian regulator should have a say over what people elsewhere can see.

After the 2024 stabbing of Bishop Mar Mari Emmanuel, X blocked the footage for ordinary Australian users. eSafety demanded more because Australians could still find it through a virtual private network. The result sought by the Commissioner would have required X to hide the material worldwide. The Federal Court refused to extend the injunction, and eSafety later discontinued the case. The episode was a warning that the regulator’s ambitions did not stop at our shoreline.

Then there is Telegram. The eSafety Commissioner has taken the messaging platform to the Federal Court, accusing it of failing to remove pro-terror material and videos of extreme violence. The allegations include terrorist executions and footage from the Christchurch and Buffalo mass shootings.

That footage that would otherwise be broadcast on the nightly news can't be published on the internet raises questions. Nonetheless, Telegram apparently has been trying to comply with the law and says it has made extensive efforts to fight terrorist content and will contest the case. Ultimately, the Federal Court will decide whether it broke Australian law.

But the concern is how much authority one regulator is accumulating. Telegram has more than a billion users worldwide, yet Australia is threatening it with penalties of up to $54.6 million under laws written in Canberra. Inman Grant has also reportedly floated the idea of asking a court to block access to Telegram in Australia if the company were found liable and refused to pay.

There is a perfectly reasonable argument that any company offering a service to Australians should obey Australian laws within Australia. But when those laws are trying to enforce censorship, I, for one, am thankful that foreign companies are thumbing their nose at the government. But that's about what happens in Australia. Trying to dictate what the company does overseas, or compelling foreign employees to surrender records, goes much further.

The worst material on the internet makes this expansion of power easier to sell. Few people will object when the target is an Islamic State execution video. It's grotesque, and so people want it banned. People also want free beer on Fridays but they don't get it. My thoughts are that people should actually be able to watch ISIS murdering folks to get a sense of the gravity of the situation the world is facing. That being said, the bigger problem is that wider powers remain on the books after an immediate horror has passed, ready to be applied to the next category of “harm”. And who defines harm?

The Institute of Public Affairs has linked this bill to Labor’s proposed digital duty of care. That policy would cover far more than the major social media companies. Messaging services, apps, websites, and other online providers could be required to take “reasonable steps” to prevent a range of broadly defined harms.

We have already seen what Labor’s idea of “reasonable steps” looks like. The regulator decides what it expects, demands the records, and can take the matter to court under the threat of enormous penalties.

Companies will protect themselves. If excluding a lawful adult costs nothing while admitting an underage user could contribute to a $109 million penalty, the safer business decision is obvious. Anyone whose age cannot be established with enough certainty will be subjected to more checks or locked out.

That means more facial-age estimation, more identity documents, more behavioural analysis, and greater use of outside verification companies. People who simply want to read the news, follow a political discussion, or watch a video will increasingly have to prove that they are adults.

Ruinous penalties can create an online identity system without Parliament ever passing a law that explicitly orders every Australian to identify themselves. Platforms will collect the information because the government has made it too dangerous not to.

Once those systems exist, they can be extended. The same age checks can be used for adult material, online marketplaces, and any other service later declared harmful or unsuitable. The machinery will already be sitting there.

Australia is poorly equipped to resist this sort of control. Our Constitution contains no express protection for freedom of speech. The High Court has recognised an implied freedom of political communication, but that is a restriction on government power rather than a personal right held by each Australian.

We have nothing equivalent to the USA's First Amendment. That creates a genuine conflict when an Australian regulator issues demands that affect companies, employees, and materials in the United States. Speech protected there does not lose that protection merely because Julie Inman Grant dislikes Australians being able to access it.

The Senate committee examining this bill is due to report on 25 August. It should reject the legislation.

If senators will not do that, compulsory demands to third parties should at least require independent judicial approval. The law needs a clear Australian territorial connection, proper protection for confidential and legally privileged material, strict limits on access to personal information, and an explicit safeguard preventing ordinary users from being dragged into eSafety investigations.

Labor should also be made to show that its original ban has improved children’s lives before it receives any more power to enforce it. So far, eSafety’s own figures show children continuing to use social media while more parents remain unaware of what they are doing.

Parents need useful tools and honest information. Platforms should face consequences when they knowingly ignore criminal material or deliberately push harmful content at children. Neither of those goals requires handing one unelected regulator access to private records across the world.

Musk has his own reasons for fighting this bill. I have no particular desire to defend a billionaire or any of the giant technology companies. The principle matters more than the personalities involved.

Government agencies rarely return powers once Parliament has handed them over. With the under-16 ban failing and eSafety already testing the limits of its reach, giving Julie Inman Grant even more authority would be reckless.

“Necessity is the plea for every infringement of human freedom. It is the argument of tyrants; it is the creed of slaves.”
– William Pitt the Younger