The Government wants to filter the internet. Is that a good idea?
Thursday, 23 July 2020
The Government plans to filter out the worst excesses of the internet. Experts think it will have almost no impact, other than introducing a series of grave risks. National Correspondent Katie Kenny reports.
Few people would argue for the right to watch and share a video of mass murder on social media. Or for images of child sexual abuse to be available online. And that’s probably why the Films, Videos, and Publications Classification (Urgent Interim Classification of Publications and Prevention of Online Harm) Amendment Bill hasn't caused too much of a stir.
But the proposed legislation raises important questions about the extent to which the Government should be able to control what we see and do online. Particularly when the content is less obvious than terrorism and child abuse.
In short, the Bill would allow the Department of Internal Affairs (DIA) to operate an electronic system to block public access to objectionable online publications – a web filter. The filter could prevent access to an entire website, or part of a website.
Details beyond that are vague. The regulations dictating the filter’s design and operation would be written by DIA. Internet service providers, technical experts, online content hosts and the public would all have a say in it. And there would be a review and appeal process.
**READ MORE:
* Heads butt over new bill’s internet powers
* Christchurch attacks: Facebook removes 4.5m pieces of violating content in just six months
* Internet service providers block video game inspired by terrorist attack
**
Officials have noted the risk that “civil society views the changes as unduly limiting freedom of expression” but that there was increasing evidence about the psychological effects of online exposure to mass violence and extremism.
Internal Affairs Minister Tracey Martin, when introducing the Bill to Parliament on May 26, said: “This is about protecting New Zealanders from harmful content they can be exposed to on their everyday social media feeds.”
The catalyst
On March 15, 2019, as a white supremacist shot dead 51 worshippers in two Christchurch mosques, a camera strapped to his helmet captured the massacre and broadcast it live on Facebook. The 17-minute clip was viewed about 4000 times before the social network removed it, but by then it had spread across the internet. Within the next 24 hours, Facebook removed 1.5 million copies. YouTube at one point was removing one copy per second.
The country’s biggest telecommunications providers moved to block access to domains hosting the material. After blocking a domain, internet providers contacted the hosts and asked them to remove the footage of the killings to have their domain unblocked.
The Christchurch attack was later described as “social media-fueled terrorism”. Prime Minister Jacinda Ardern vowed to reform social media and curb “the horrifying new trend”.
Two months later, at a summit in Paris, hosted by Ardern and French President Emmanuel Macron, the Christchurch Call to Action was announced. The pledge asks nations and private companies to make changes to prevent the posting of terrorist content online. So far, 48 countries, the European Commission, two international organisations, and eight technology companies have signed it.
New Zealand’s filtering legislation would support the Call, Martin has said.
But the Call also rests on the conviction that a free, open and secure internet offers “extraordinary benefits to society”. And that doesn’t sit well with filtering it, says InternetNZ chief executive Jordan Carter.
A decades-old debate
Some will remember a similar debate more than a decade ago, when DIA announced it would block access to blacklisted websites that hosted child sex abuse material. Known as the Digital Child Exploitation Filtering System, it’s a voluntary system that sits outside the law.
When a user in New Zealand tries to access one of the thousands of web addresses listed on DIA’s server, the internet provider, such as Spark, or Vodafone, will automatically divert the user to a notice saying the site they’re trying to reach has been banned. The notice has information about counselling and how to anonymously appeal the ban if the user believes it’s unfair.
The list of blocked addresses is based on referrals to and investigations by DIA’s digital safety team.
About 85 per cent of New Zealand internet providers comply with the filtering system. It’s generally considered to work well except for when it accidentally took down Google in New Zealand. (In 2013, DIA confirmed it filtered some sites hosted by Google, causing problems for both the filter and some internet users.)
“We don’t think the Government should give itself the power to filter web content,” Carter says. “We don’t think that’s consistent with having the free, open and secure internet described at the Paris summit.”
His issue isn’t with filters, it’s with the state mandating one. “People use filters all the time. Corporate networks, school networks, we have a filter we offer commercially to blocking phishing and malware.
“It’s problematic when you combine that technology with state power. At the most extreme end, you have countries that do whole internet shutdowns. Governments have just flicked off the switch.”
If it were voluntary, like the existing child exploitation filter, Carter would be more relaxed about it.
If future governments tried to increase the scope of the filter, internet providers could walk away, he says.
“If you want to offer the filter, go for your life, but don’t put it into law.”
Martin has said she wants to protect Kiwis from harmful content on their everyday feeds, but the likes of Facebook, YouTube and Twitter are already incentivised to remove content that breaches their terms and conditions as quickly as possible. Filtering, then, isn’t likely to have much of an effect on the big platforms, Carter reckons.
A filter also isn’t going to stop people who are determined to access encrypted content on dark web sites, Carter says, or anyone who’s actively seeking content through tools such as virtual private networks.
“So you’re looking at the cases in between. News sites. Random blogs. The filtering approach might help with that.”
A slippery slope
In his submission to DIA, Dr Andrew Chen, research fellow at the University of Auckland's think tank Koi Tū, The Centre for Informed Futures, recommended the web filter be a separate piece of legislation.
The Bill’s other proposals – allowing the chief censor to more quickly notify the public of objectionable content, making the live-streaming of objectionable content a criminal offence, giving the Government the ability to issue take-down notices to hosts of harmful content, and bringing social media companies within the scope of current laws – were fine, he says.
“Normally I don’t like slippery slope arguments … but if we have a filter for extremist content, would we then want a filter for name suppression?”
Like Carter, Chen says those who really want to get around the filter will be able to do so.
Ten years ago, when the DIA introduced the child exploitation filter, “they said that was it”, Chen recalls.
The Act’s current definition is relatively narrow, he says, and the chief censor does a good job of applying it.
But he worries about scope creep. “Just because the current government defines it in one way, doesn’t mean a future government wouldn’t go further. It’s hard to undo these things. They only ever go in one direction.”
Content moderation is a global issue. Every big technology platform is wrestling with the question of how responsible it should act for the material people share.
If the technology giants are struggling, “I’m not sure how DIA is going to do it”, says Tom Barraclough, director of Brainbox Institute, a research centre for emerging technologies and law.
“If we’re going to be normalising the Government’s ability to filter what we see online, we need to be proceeding down that road very carefully.”
But internationally, the concept of ‘one internet’ is already eroding. “We’ve started developing separate, national versions of the internet. For example, China has a separate internet, and other countries have national firewalls, too.
“In New Zealand, we have a lot of trust in our government to not go down that route. But I think we’re entitled to have more legal remedies.”
One of his main concerns is the huge amount of work being delegated to regulations which won’t go through Parliament. “I think that’s not appropriate, given the power at stake.
“We’re about to adopt a legal regime that allows algorithms to police content online.
“We’re putting a lot of reliance on DIA and the chief censor to sort out these tricky questions. If Parliament is saying, ‘this is the right thing to do,’ it’s the MPs who should be owning this and having the hard discussions with New Zealand.”
Potential backlash
Barraclough worries about the social fallout, too. Even with the right checks and balances, if the public doesn’t have that perception, there’ll be a backlash. In the worst case scenario, it could drive more people towards anti-state, conspiratorial agendas.
Barraclough, along with his colleague, Curtis Barns, has produced some of the country’s pre-eminent research on legal and institutional responses to the potential impact of synthetic media and disinformation in general.
“When it comes to talking about disinformation and ‘deep fakes’, one of the core elements of a strategic disinformation campaign is to lower people’s trust in institutions and each other.”
Sometimes, someone’s attempt to hide something online can cause it to go viral. This phenomenon is known as the Streisand Effect, named after Barbra Streisand’s unsuccessful legal attempts to suppress the publication of photos of her Malibu house. The campaign only resulted in more online attention.
Any public reporting on extremism has to be a careful balance between telling important stories and unwittingly amplifying the messages of hate groups.
Preventing access to material only would only fuel suspicion, Barraclough says, perhaps fuelling desire for access.
“We’ll absolutely see some people get the wrong end of the stick and say, the Government is policing what they see online, they’re hiding the truth, and so on.”
The antidote was transparency, legal accountability, and democratic oversight, but the draft legislation suggested “a lot of hard questions are being kicked down the road”.
Ben Elley, lead researcher at Independent Research Solutions, says while banning the video and manifesto produced by the March 15 shooter was a good move, it’s turned them into somewhat “valuable commodities”.
He’s observed fringe groups relish in their content being banned.
In reality, filtering objectionable content isn’t going to stop extreme communities from sharing it, Elley says. But it could stop people from being inadvertently exposed to it via mainstream websites.
While someone is unlikely to be radicalised from stumbling across a single document, there’s a risk in “gradually stumbling” down a radical rabbit hole.
“People will gradually move from one Facebook group to another, and at the same time become more alienated from society.
“It tends to be more about communities than documents.”
Rather than silence, add more voices
Anjum Rahman, of the Inclusive Aotearoa Collective, agrees. “Often it’s not the single post that’s the problem, it’s the fact there’s a whole lot.”
But the answer isn’t necessarily in trying to silence people, she says. Rather, it’s in empowering the speech of vulnerable groups – implementing policies that allow them to speak back and balance the platform.
There are already practical solutions at our fingertips, she says. She points to New Zealand’s mainstream media and the lack of diversity among opinion columnists and cartoonists.
While she agrees in theory that content promoting the sort of violence witnessed in Christchurch should be blocked, she’s concerned a national filter could end up disempowering those who already struggle to have their voices heard.
“When you have tight laws and regulations, it’s usually minority voices that have those used against them, because they have less power in society.”
Any web filter needs to be co-designed with these groups, she says. “You need people in the room who can speak to how they’re impacted.”
At the end of last year, DIA led seven consulation workshops across the country. Attendees included representatives from the major online content hosts, internet providers, community groups and civil society organisations. About 200 people attended.
There was no formal consultation with iwi or hapū before policy decisions were made. DIA invited stakeholders to submit feedback; 17 submissions were received.
Chris Cormack, a technical lead and kaihuawaere matihiko/digital facilitator at Catalyst IT, puts it simply: “I don’t think a filter is going to solve what [the Government] says it’s going to solve.”
The potential harms of inadvertently blocking content that shouldn’t be blocked will outweigh the positives, he says.
Despite the promise of extensive consultation, Cormack isn’t convinced that will happen. “It’s telling that we’re this far along, and that hasn’t happened so far.”
On drawing a line
Since taking on the role in 2017, chief censor David Shanks has been repeatedly thrust into the spotlight. First, over the controversial Netflix series, 13 Reasons Why. More recently, over violent extremist content.
While the Government is keen on sending more money and power his way, Shanks hasn’t been shy in raising concerns about the long-term shape of his office.
He doesn’t think the Bill is perfect, but it has the potential to make a “very significant difference”.
“Just as we already take practical steps to limit the distribution of child abuse images, we should act to restrict the spread of illegal material that promotes terrorism and celebrates mass murder,” he says.
The request from internet providers for clearer guidelines for blocking content is reasonable, he adds.
“The fact that some dedicated criminals might be able to work around a filter doesn’t change the fact that filters will stop this material from getting through to many people. And I think it is very achievable to get the balance right between protecting people from deadly extremist material, while protecting every New Zealander’s freedoms.”
Dr Marcin Betkier, a law lecturer at Victoria University of Wellington, says the days of a free and friendly internet have been and gone.
“It’s evolved into a structure which is non-transparent for individuals and governments, in which power is held by big companies that mediate interactions between people.
“They have monetary incentives that don’t necessarily support individual or societal wellbeing. And we can’t rely on them to self-regulate.”
He points out the Bill proposes a filter that only monitors web pages, not communications.
In other countries, such as Poland, where he used to live, the legislation would be considered too permissive, too powerful. But here? “I think the Government is more controlled and scrutinised. The political culture is more trusted and open.”
However, he wonders how much notice social media giants will take. Facebook has previously said New Zealand laws don’t apply to it. Google has ignored our suppression orders.
He’d also like to see a case made for gathering more data around the proposed legislation’s effectiveness. “How many people were viewing or downloading content? There’s a difference between something being on Facebook and being seen by millions of people, versus a few people on a small site.”
Like Shanks, he believes it’s better to implement an imperfect solution for tackling harmful content online, than allow things to continue as they are.
“We have to draw a line. We can’t sit here and say, ‘it’s impossible, so let’s do nothing.’”