On the proposed regulation of social media platforms
In the wake of the Christchurch Mosques terror attacks on 15 March 2019, New Zealand Prime Minister Jacinda Ardern has said that social media platforms must take responsibility for the material that they distribute, that they are the publisher, not the postman. More specifically,
We will look at the role social media played and what steps we can take, including on the international stage, and in unison with our partners. There is no question that ideas and language of division and hate have existed for decades, but the form of distribution, the tools of organisation, they are new. We cannot sit back and accept these platforms just exist, that what is said on them is not the responsibility of the place they are published. They are the publisher, not just the postman. There cannot be a case of all profit, no responsibility.
This article examines the spectrum between postie and publisher, and explores the nature and locus of accountability for what we say and do online.
Note: I’ll be using the gender-neutral term “postie” rather than “postman” in the remainder of this article.
People have been hating each other since the beginning of time – even the bible takes less than 100 verses to get into the story Cain and Abel, the first murder.
Back in the day when information was distributed through physical media, publishing information took time, energy, significant financial resources, and organisation. Governments could easily regulate and control publishers, as there were relatively few of them. But the Internet changed all that – anyone can now disseminate information to a global audience instantaneously through a wide variety of channels in different jurisdictions.
How can we approach making those who disseminate information responsible and accountable?
First, some general principles:
- Your rights and responsibilities should be the same, whether online or in physical space. There should be no difference in my right to make a television broadcast versus my right to make an online broadcast.
- Freedom of expression is a cornerstone right in a functioning democracy. This freedom has limits, and does not extend to incitement, publishing of private information (eg doxing), libel, or other communication in which people are harmed. In New Zealand, we have censorship laws to protect people from publications containing objectionable material that is “injurious to the public good”
- Private conversations should not be subject to government surveillance. I should have a right to exchange confidential information with individuals and institutions such as my wife, my lawyer, or my bank without fear of surveillance, in my bedroom, my office, a restaurant, or online.
- The network itself is just a tool to get messages from one place to another. The Internet’s superpowers are being Direct, Open, Accessible, and Free (see my TEDx talk for more on this). The Internet’s design point was to be able to continue to effectively function even if most of it were to be taken out by a nuclear strike – it views censorship at the network level as a bug, and will route around it. Regulating the network itself is neither desirable nor practical, in possible contrast to regulating the services that run on the network.
You could say that the network itself is like the postie that delivers messages from one place to another, but the social media services that run on the network are more like publishers. But the question of postie vs publisher does not generally have a binary yes-or-no answer, and services fall into a spectrum.
A number of different factors contribute to where a service falls on the postie-publisher spectrum, including:
- Distribution: is the service mainly used for distributing messages 1:1 or 1:several (postie) or 1:many (publisher)?
- Ephemerality: do messages disappear from the service once they’ve been delivered (postie) or do they stick around (publisher)?
- Content inspection: does the service just pass the message through (postie) or inspect the content of the message and potentially take action depending on the content (publisher)?
- Revenue model: does the service make money per transaction (postie) or do they monetise based on content or viewer attributes (publisher)?
Using this model, let’s look at a few specific examples:
Email (SMTP) is generally used for 1:1 or 1:several, messages disappear once they have been delivered, servers do not generally inspect content (although some services try to classify SPAM), and is generally bundled with existing services, noting that some services eg Gmail advertises to you based on the content of your messages [one of the many reasons I don’t use Gmail]. Email is pretty clearly on the Postie end of the spectrum.
Signal, WhatsApp, Viber, and other direct messaging clients are similar to email. Their protocols use end-to-end encryption so that the platforms can’t read the messages as they pass through. But they could analyse the metadata if they wanted to, which has caused some concern. Again, they don’t advertise to you based on content, so they too land on the Postie end of the spectrum.
Slack, HipChat, Jabber and other workplace messaging clients are typically used within specific business communities, and used for 1:1 and 1:several communications, although they do have the ability to be used for 1:many. Messages can hang around for a very long time depending on the subscription plan. The services do not take much action (other than for example providing link previews) based on content, and do not have advertisements. These services are still mainly Postie, but starting to move slightly toward Publisher.
Reddit, Hacker News, 4chan, and other message board sites are generally used to create communities of people to discuss specific topics. They are used almost exclusively for 1:many communications, and messages hang around indefinitely. Some of these services have explicit editors, and some have message upvoting/downvoting which is effectively crowdsourced editorial control which helps them create and support filter bubbles. Reddit and 4chan’s revenue models are based on advertising, whereas Hacker News is a free service. These services look more like Publishers than Posties.
Kiwiblog, The Standard, Whale Oil, and other blogs are 1:many, and the messages stick around forever. Some blogs do not monetise, others monetise through adverts and sponsored content. They are Publishers, self-funded or otherwise.
WordPress, Medium, Wix, Weebly, Squarespace, and other content authoring and distribution platforms provide infrastructure for bloggers. Even more than the bloggers themselves, they are Publishers. Many of them provide free accounts which monetise through content-specific advertising.
Twitter, Facebook, Instagram, YouTube and other social media platforms fit all of the criteria above. They are publishers.
Publisher sites like Stuff, The Herald, Newshub, and Radio New Zealand, are very clearly publishers.
But what can we do?
We already have some good tools at our disposal by way of our existing legal framework and specific legislation. I’m no lawyer, but it seems clear to me that the Films, Videos, and Publications Classification Act 1993 provides a good starting point. But there are two significant problems.
Speed: As we saw with the spread of the objectionable video of the Christchurch shooter, objectionable material can spread very quickly. Once it is out, it is very difficult to control. I don’t believe it is realistic to expect any platform to be able to function at speed and scale at the same time as preventing at least some distribution of objectionable material, sometimes. The alternative of having a human check each piece of uploaded content before releasing it for “publication” would slow the system down so much that it would no longer be workable.
Relying on algorithms to do the job is fraught too: consider the case of Good Bitches Baking, a nonprofit network of people who want to show kindness to those in their communities who are having a tough time. They do this by baking delicious treats. And yet whenever they attempt to post on Facebook, they are blocked because of their name.
Dear @facebook What, exactly, is offensive about these @GoodBitchesBake tees? They literally say “before all else, be kind” in English and Te Reo Maori. So, unless you find us using both official languages of our country offensive, I’m stumped. pic.twitter.com/Vcu5Y6Al0T— Marie (@LadyInDread) November 22, 2018
Jurisdiction: As a small country of only 5m people, no matter what we do, it’s going to be hard to provide teeth to any attempt to regulate the big multinational platforms. We’ll need to work together with friendly larger and like-minded countries. We don’t want to create a “Great Firewall of New Zealand”, and even if we did, it would be of limited value.
Facebook claims that the video of the Christchurch shooter was only downloaded a couple of hundred times before it was taken down, but by that stage it had already been downloaded from Facebook and uploaded to 4chan, 8chan, and other platforms. Spark and other ISPs took their own initiative to block some sites on the day, but these blocks have since been removed.
Confirmed. NZ ISPs are blocking access to #Liveleak, 4chan, 8ch, a certain farm, Mixtape, Mega, and many other sites that are not complying immediately with the takedown orders. https://t.co/eaBJvvRc5E— KiTA (@eldarmark) March 15, 2019
Even @Bitchute is removing the video mirrors.
cc: @nickmon1112 @wikileaks
Chasing content round in this way is akin to playing whack-a-mole – whenever you block the content, you find it pop up somewhere else. The collateral damage to legitimate and legal content on these platforms seems disproportionate. And the ultimate destination is the Dark Web. Do we really want to block that too? There will always be objectionable material available on the Internet, outside of anyone’s control.
And what do you do with people who want to be misinformed? Copernicus’s theory that the earth revolved around the sun and not the other way round was considered dangerous misinformation when it was released. Perhaps it only matters when people are harmed by the misinformation. I’ve said elsewhere that separately legislating hate speech on the Internet is a bad idea, and drawing the line between hate and harm is difficult.
These are hard problems to solve, and even more difficult to solve quickly and in isolation.
Jacinda is on the right track – social media platforms should be more accountable than they are at present. But I urge the New Zealand Government to take a considered approach, which may take some time to get right, and work hard to balance public safety versus freedom of expression, and understand the nuanced spectrum between posties and publishers. We don’t want “napkin doodles” being rushed through Parliament. New Zealand has achieved a good balance in the past, and I believe we can do so in the future.