Australia Confirms Censorship Plans, Tells Fibs On The Filtering Trial

The Australian Government today confirmed that it would proceed with broadscale internet censorship in Australia following a trial into ISP based internet filtering.

Delivered pre-Christmas so as to minimize debate on the plan, Reichsminister for censorship Stephen Conroy spun the decision by selectively quoting parts of the trial report. Conroy claimed among other things that banned material “can be done with 100 per cent accuracy and negligible impact on internet speed” and that the filter would apply to all RC (Refused Classification) content.

It’s an interesting line, because the reports findings don’t actually say that the filtering of RC content is either 100% accurate nor would it necessarily have a “negligible” affect on internet speeds (we’ll get to “negligible” in a moment.)

Filtering accuracy: an inconvenient truth

The study asked the nine ISP’s to trial several forms of internet filtering. The first test was based on the flawed ACMA blacklist of approximately 2,000 sites, which as we know from earlier in the year blocks dentists, poker sites, and other legal sites along with illegal sites. The second was based on a broader “child safe” filter which attempted to filter more sites (although the exact figure was not disclosed.)

The tests found that 100% accuracy was obtained with the ACMA blacklist only, a list of 2,000 odd sites that would only be a small sample of sites blocked under the scheme.

The tests found that when the list was expanded to the bigger child safe list, that accuracy dropped to between 78.8% and 84.6%.

On those results, the study claimed “Enex considers it unlikely that any filter vendor would achieve 100 percent blocking of the URLs inappropriate for children without significant over-blocking of the innocuous URLs.”

Notably a full list under the implementation of censorship would be significantly larger than what is currently in the blacklist, a point even Conroy concedes. Depending on the estimation, the real world list could total anywhere from 20,000 sites to hundreds of thousands. So how can you claim a future filter is 100% accurate when the report itself notes that accuracy will never be 100% on a bigger list?

What Is Actually Being Filtered?

While Stephen Conroy claims the filter will apply to RC content, the report didn’t just consider RC content, and actually goes as far as noting that the blacklist includes “prohibited” content.

The Online Content Scheme (the Scheme) introduced in 2000 under the Broadcasting Services Act 1992 (Cth) (the BSA), regulates content on the internet. The Scheme is contained chiefly in Schedules 5 and 7 of the BSA. Under the BSA, ‘prohibited content’ and ‘potential prohibited content’ include content that has been classified or is likely to be classified Refused Classification under the National Classification Scheme.

Better still, the report actually discusses how prohibited content will be added to the blacklist:

Where content is not hosted in Australia and is prohibited, ACMA will notify the content to the suppliers of approved filters so that access to the content using such filters can be blocked.

It should be noted that prohibited content in Australia is broader that RC alone; content may be prohibited if it doesn’t adequately have age verification for example.

So which is it Senator Conroy? And if the study was accurate, why wasn’t it testing RC content alone? At the same time if the report is based on Government specs, why does it talk about prohibited content?

Define Negligible?

The most blatant sign of bias in the terms of the study is how the Government, and by extension the study decided to define negligible. The report defines it as such:

Negligible impact on network performance

10% and below: individual service performance impact is negligible to the end-user. It would be difficult for the test to distinguish the impact of the filter from any other factor potentially affecting network performance.

Yes, a drop in speed of 9% is considered negligible in a country where internet speeds are at times third world.

But here’s where it gets better: the findings on speed were all over the place, with some results even questioned by the study, like the ISP who reported that speeds were 17.32% FASTER with the filter (WTF doesn’t begin to explain it.)

Some highlights: the biggest decrease in speed is reported to be 44.15%, with other ISP’s reporting speed cuts in various tests of 36.45% and 27.2%. A range of results came in at 0-10%, with some others reporting small increases on some tests (where a proxy to deliver some data was used.) But even those reporting increases of speeds in some tests also reported decreases in others, for example participant 9 reported a small increase in streaming speeds, but a 10% decline in web page speeds.

While you can concede that the average decreased speeds aren’t as bad as some may have feared, to suggest that the results were negligible is grossly disingenuous; Wikipedia defines the word as referring “to the quantities so small that they can be ignored (neglected) when studying the larger effect.”

Circumvention, and a ban on Proxies?

The report also details the circumvention of blocked sites. It notes in its summary that “A technically competent user could, if they wished, circumvent the filtering technology.” The stats: on the ACMA blacklist, circumvention was blocked at a range of a pathetic 8.1% to a high of only 16.2%. The broader list (with the tests of the broader filter being opt-in by basically those wanting to block stuff from their kiddies to begin with) saw results from 37.8% to 94.5%.

The note there is that it’s the stuff on the blacklist (in theory currently, and the expanded version) which is what the Government really wants to block (dentists with kiddie porn or something like that.) That the best result of circumvention blocking was 16.2% on the hard stuff really highlights what a sad and pathetic joke the filter is; those who want to look at this stuff will still be able to and the filter isn’t going to stop them, and that’s before we consider that the report notes that they can’t block IM or P2P.

But the other strange assumption in the report is that circumvention is only in the domain of a “technically competent user,” a suggestion that implies that it is somehow difficult to use a proxy to bypass the filter. The problem is that it’s not. I ran proxies a few years back for a brief time, and the most popular site accessed via my proxies (this is before Facebook got big) was MySpace: proxies are popularly used by non-technical people to bypass work and school filters, often by kiddies, and you don’t need to be a genius to use one; type proxy into Google, or visit a directory site like

Likewise the report then seems to accept that proxies aren’t that hard to use to some degree, and this is where it gets into scary territory: the report suggests that proxies should be banned:

Filtering of additional categories of content enabled ISPs to implement measures which made some common circumvention techniques difficult. For example, a third party website which hides the origin of the requested content (proxy site) can be included in a wider list of URLs to be blocked.

and on the broader (not ACMA blacklist) test that delivered better anti-circumvention results, the study notes that proxies were being filtered as part of the test:

The testing for circumvention generally indicates that filtering of additional categories of content, enabled ISPs to implement measures which made some circumvention techniques more difficult to use. For example many commercial lists have a “proxy” category; proxies are a common form of filter circumvention.

Anyone for adult computer games?

Conroy claims that adult computer games will not be initially included in the filter as the Australian Government has started “a public consultation process into whether there should be an R18+ classification category for computer games.

One small problem with that: censorship power in Australia resides with the States. All the states have to agree to support an R18+ rating for computer games for any Federal application, and South Australia isn’t about to change its mind. WA has also banned possession of RC computer games as well, so there’s no surety that the WA Government would support a change either.


It’s a sad day for freedom of speech in Australia when the Government delivers a slanted report that when you actually read it doesn’t back everything it is claiming. It’s not dissimilar to Iran: you don’t get the result you want so you ignore the results you don’t want, you stuff the boxes in other places, and you deliver a result that was always predetermined to begin with.

Conroy claims again that the filter is all about kiddie porn (which is already illegal to view and host anyway) but at the same time will ban euthanasia and abortion sites, along with adult computer games as well. The dark clouds of totalitarianism are descending on Australia; remember, history shows that Governments who start on the road to censorship usually expand the regime with time. The can is open now, and who knows when the madness might end.