Australia's Online Censorship Push: Who is Funding the Campaign to Control Speech?

Recent investigative reporting has examined the funding and institutional relationships surrounding Australia's online regulation debate. Rather than repeat or adjudicate those claims, this article concentrates on the publicly available research used to justify the proposed legislation.

The internet, however, is still open, for a while before total censorship. For the next few days at least you can read the original investigation yourself and decide what you make of it:

Michael Nayna and Andrew Lowenthal in Public, 29 October 2026: https://substack.com/home/post/p-222429140

Related commentary from Rebekah Barnett:
https://news.rebekahbarnett.com.au/p/evidence-free-social-engineering

Have a Captain Cook at those.

What can be discussed on the public record, without naming anyone, is the quality of the evidence being used to justify the Digital Duty of Care proposals and the amount of government money flowing to some of the advocacy groups involved.

The widely repeated claim that social-media algorithms show young boys encounter misogynistic content within 23 minutes, comes from experimental accounts (some deliberately seeded toward manosphere material). It does not demonstrate that the average Australian boy encounters violent misogyny after 23 minutes of ordinary use. The study's own finding is toxic content at 23 minutes and manosphere content at 26, and the widely repeated "average of 23 minutes" is wrong twice: ten experimental accounts cannot establish a representative population average, and 23 was a ceiling. The gap between the study design and the slogan is material.

Similarly, a frequently cited 99 per cent figure on women experiencing "abusive behaviour" on dating platforms in a three-month period, rests on a broad definition that folds in serious conduct alongside more subjective categories such as love-bombing and negging. Aggregating them produces a striking headline; it does not isolate the prevalence of threats, stalking or image-based abuse; 95% psychological/emotional, ~59% fear-based controlling behaviour including threats of physical harm. "At least one harmful behaviour" is doing enormous work in a 99% figure.

Australia already has criminal laws covering threats, harassment, stalking, sexual exploitation and the non-consensual distribution of intimate images, plus existing eSafety powers. The open question with a broad "duty of care" framed around foreseeable harm is whether platform incentives will stay inside those established categories or expand into lawful but contested speech. That design question, and the strength of the research used to answer it, can be examined on the public record without needing to identify any particular philanthropist.

The original reporting is still up. Read it while you can and form your own view.

https://substack.com/home/post/p-222429140

https://news.rebekahbarnett.com.au/p/evidence-free-social-engineering

"The evidence base for this intervention is a house of cards.

We've seen this before with the government's failed 'misinformation bill.' The government claimed that online misinformation was an "increasing" (though unquantifiable) concern. Yet the government's own report justifying the bill contained misinformation, and the two studies it relied on miscategorised true information as 'misinformation' simply because it contradicted government advice at the time.

It's depressing to discover, over and over, that the people in charge are running an evidence-free fiasco in cahoots with myopic academic centres and … with a penchant for social engineering.

Covid was the big eye-opener on this front, but once you see it there, you can't help but see it everywhere thanks to good researchers and writers finding out the what and how of it and packaging it up for aghast readers."

-Rebekah Barnett