Over-Filtering The Internet

My point of view on all of this: Edgy content always finds a way

Late last year, when I heard about the passage of Australia’s ban on social media for kids under the age of 16, I described it as “kayfabe”—i.e. pretending to take action on a politically valuable issue.

Someone challenged me on the premise, stating they didn’t understand my opposition to it, saying something along the lines of “The idea itself isn’t bad, it’s just not very enforceable.”

Let me be clear: The idea itself is bad, in my estimation. It is one of many examples of a law that exists to make a statement rather than something that truly solves a problem. And I think the decision to include YouTube as part of the ban underlines this especially.

And I get the motivation—it is an effective way to look like you’re doing something as a government. But it is a real slippery slope kind of situation that often introduces a bunch of dangerous stuff along with it.

And the truth is, it often doesn’t work. Kids are smart. They can work around stuff like this—and have, time and time again. (A teenager with a desire to work around a restriction is a hell of a pentester.) But we’re still stuck with all the bad junk that gets dragged in with these seemingly well-intentioned laws.

Taylor Lorenz has been on top of this beat in a big way in recent weeks, writing multiple articles about it in The Guardian and elsewhere. Recently, in her newsletter User Mag, she brought in a trio of researchers—Cynthia Conti-Cook of the Collaborative Research Center for Resilience, ACLU lawyer Rebecca Williams, and activist Pratika Katiyar, also of the ALCU.

(Side note: She brought User Mag to Patreon recently, which, as someone who has a thing against Substack, I appreciate.)

The guest post has many great parts, but I especially think this point about digital ID systems is especially important:

Age verification laws don’t just fail to protect children; the digital ID systems that enforce them fundamentally change how everyone accesses the Internet and will cause real harm.

Once the infrastructures of unique identification are constructed, it won’t only be advertisers profiling and microtargeting customers but states profiling and microtargeting any political opponents or other public enemies-du-jour. That is in addition to chilling free speech, restricting access to vital health and identity resources, disproportionately impacting historically oppressed communities, and forcing users to hand over sensitive biometric or government ID data to unaccountable third parties.

We must stop the harms caused by digital ID and paternalistic laws like age verification. We can resist with daily acts of resistance, organize, legislate real protections, and litigate.

That’s the key thing for me. It always starts with protection mechanisms intended for one audience, but it changes the way every audience uses the internet. It brings in a new class of companies and motivations, and those motivations can be quite dangerous if not correctly managed.

And we have distinct proof of this in the form of the filtering tools and legislative tactics we’ve seen in the past.

How online filtering created a new front in the culture wars

While there are plenty of examples of networked technology before the late ’90s—the Free-Net, for one—things started to pick up after the web became a thing. It started appearing in common settings; think schools, cybercafes, offices, and so on. Things had reached scale.

And this created a need for a market for digital filters, which were intended to serve a role not that dissimilar to the V-Chip—that blocked the bad stuff from being accessed online, while allowing most of the good. This sounded good in theory, but ultimately, the filters weren’t very good. And honestly, the “I know it when I see it” approach to indecency and obscenity, as famously defined by Supreme Court Justice Potter Stewart in the 1964 case Jacobellis v. Ohio, breaks down online. When new webpages are being produced by the thousands and even millions each day, you can’t possibly see everything, and suddenly, it becomes a matter of grappling with a whole lot of different standards of what’s safe and what’s not.

In other words, it was a First Amendment issue, and a knotty one at that. Installing the V-Chip in TVs? Compared to content filters on the internet, relatively painless—as soon as Al Gore was convinced, it became downright easy to make the case for it. It empowered parents without actually blocking anything for folks who didn’t want to use it. While some broadcasters might have felt frustrated by a decision that could affect their ad revenue, it was easy for folks who weren’t in the target audience to ignore.

But internet filtering software had many more variables. The internet didn’t have a standardized rating system like television or movies. Anyone could create anything on it—and everyone did. Filtering required much more room for edge cases, such as that of the British city of Scunthorpe.

Naturally, this issue becomes a challenge as Net Nanny, WebSense, and so on, filter things with differing sets of standards. That means you’re basically controlling content based on someone else’s opinion of what’s decent and what’s not—as well as how often they choose to update their filters. Something, inevitably, would get through. Kids are smart.

A 1999 editorial in the Quad-City Times really nailed the problem, suggesting that by handing the job to an automated filtering app, rather than an actual person, the software created complex problems that can’t be easily sorted out by algorithm.

“Unfortunately, the software is not very refined,” the editorial board wrote. “By restricting access to sites that mention sex, for example, WebSense blocks access to sites that contain legitimate news stories on such topics as impeachment of the president.”

On the other hand, you could also argue things the other way, as Oregon librarian David Burt successfully did. Burt, concerned with the potential that children would access indecent material at the library, started up a platform called Filtering Facts. The platform used Freedom of Information Act requests filed by both Burt and a team of volunteers to highlight cases where indecent material had been accessed at libraries.

“I want to keep letting people know about the problem,” Burt told The New York Times in a 1999 article that raised his profile significantly.

The stance went against the party line of librarians at the time—that internet access wasn’t dangerous, that the freedom the internet offered was more important. Many chose not to go along with the FOIA requests, noting that existing laws prevented libraries from revealing who accessed certain kinds of information.

Burt’s research, which often took aim at traditional bodies such as the American Library Association, found support from the Family Research Council. FRC published his findings in a report titled Dangerous Access, which advocated for legal action to require filtering tools.

“The failure of many libraries to prevent these incidents combined with the demonstrated effectiveness of filtering software supports the appropriateness of legislation to require the use of filters in public libraries,” the report’s introduction stated.

Burt’s work directly helped to drive the passage of a piece of legislation called the Children’s Internet Protection Act, which required libraries and schools to install filters on computers if they wanted access to federal funding. (The latter part being key because it didn’t tie the legislation to censorship, but to funding.)

But his work actually went deeper than simply the report itself—he spoke during Congressional and regulatory hearings, before and after the passage of the law. And when the American Library Association and American Civil Liberties Union sued over the law, the Department of Justice brought him on as a consultant in the ensuing legal battle. He filed too many FOIA requests to simply let the issue go as soon as he published his report.

Dangerous Access, and Burt’s work on it, was directly cited in the Supreme Court’s 2003 decision to uphold the law. The court making its decision from the standpoint that the law did not violate the First Amendment, as the filter could be turned off for adult patrons upon their asking. The librarian who went against the grain ended up changing the law. And so, these filitering tools, some of which promoted their importance as methods to preserve free speech, ultimately ended up damaging it in one of the most important venues for free speech, the library.

It always starts with protection mechanisms intended for one audience, but it changes the way every audience uses the internet. (If it feels like I said that line already, it’s because I did.)

Burt, these days, is a consultant who previously helped Microsoft navigate the choppy waters of compliance—which makes sense, as parental controls are a compliance issue when you break it down.

Internet filtering, love it or hate it, is here to stay.