Troll Exposure
Deputy Prime Minister Barnaby Joyce reacts during Question Time in the House of Representatives at Parliament House on June 22, 2021 in Canberra, Australia. (Photo by Sam Mooy/Getty Images)

Troll Exposure

Sean Welsh
Sean Welsh
10 min read

Hell hath no fury like a trolled daughter’s father:

When multiple media houses and friends of friends from far overseas approach me and my family over a devastating, soul-destroying, career-ending lie, then the time of any person to act has arrived. It’s time to protect your daughter and so many other children from the literal filth that is the “content” of so many online platforms and their miscreant authors. My daughter is at the centre of a national scandal to which she was one of the last to be made aware—because not only is it an uncorroborated claim, it is a lie.

The father in question is Barnaby Joyce, Deputy Prime Minister of Australia, a man with some influence over the legislative agenda Down Under. He does not dignify the social media trash with further comment about his daughter. Neither shall I. In an article for the Sydney Morning Herald, Joyce says his aim is to deal with “unaccountable social media giants that give a platform to trolls and faceless cowards to engage in character assassination.” He goes on to deplore their wealth, their destruction of traditional media, and their degrading of standards of publication accountability:

Advertising revenue pours into their coffers while the traditional fourth estate of newspapers and investigative journalism move to further cost-cutting measures to stay afloat, many now as a social service rather than as an effective profit-making business. The proper inquiry into power, politics and social issues has been replaced by anonymous, slanderous, nebulous, mindless glitter.

It is hard to disagree. My Twitter feed does resemble “slanderous, nebulous, mindless glitter,” especially if I am foolish enough to read the replies to tweets posted by the sane people I tend to follow. Joyce continues:

Twitter, it is not the trolls that inspire the devastating mental health issues. The trolls don’t have a voice unless you give them one, and you do! You make money from their noise, their ambit scratchings on the back of a lavatory door. They post their character assassinations from the back of the door at the servo and you illuminate it in on a city billboard for all to see.

As they say: Don’t get mad, get even. Better yet, get mad and get even. In an attempt to hold the social media giants liable, the Australian Government plans to introduce “world-leading” legislation designed to unmask trolls. A draft of the Social Media (Anti-Trolling) Bill was released on December 1st. It arises from a preliminary ruling in the ongoing case of Fairfax vs. Voller by the High Court of Australia (our final court of appeal) on the question of “publication” as defined in Australian defamation law. The facts of Voller centre on allegedly defamatory comments made on Facebook pages run by Fairfax and News Corp. The High Court upheld decisions by NSW courts that the traditional media companies are liable for defamatory comments left on their social media pages by third parties.

The preliminary ruling on “publication” has major consequences. As the Attorney General’s explainer released with the bill puts it: “People who run normal Facebook pages can be liable for defamatory comments posted on social media even if they did not post the comments and do not know about them.” Businesses as small as a suburban café that administer a Facebook page are now potentially liable as publishers for any defamatory content posted by a passing troll while they are flat-out making coffee during the morning rush. As a result, many public figures and smaller organisations have disabled comments on their pages. This has had a chilling effect on free speech. Yet Facebook, Twitter, and YouTube, multi-billion-dollar global companies, remain “platforms” that are not liable in a defamation action. As the Attorney General Michaelia Cash said: “This is not fair and it is not right. Australians expect to be held accountable for their own actions, but shouldn't be made to pay for the actions of others that they cannot control.”

The proposed remedy is simple. As the explainer says: “Under the bill page owners will not be ‘publishers’ of the defamatory material posted on their site by third parties.” So those café owners and newspapers sharing articles on social media and permitting comments will not be regarded as publishers of the comments. Instead, parliament will overturn the High Court ruling by introducing a new law. If enacted, this would be the exact opposite of US law which currently enshrines the “platform” status of social media firms. While there have been calls from figures such as Frances Haugen and Joe Biden to repeal or reform section 230 of the Communications Decency Act, as it stands this states: “No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.”

Should the Australian bill become law, this could be the beginning of the end for social media platforms. If other nations take up the idea, they might end up becoming social media publishers instead. A major effect of the proposed bill would be to shift the burden of moderation and prompt removal of defamatory material onto the social media providers not individual page or account admins. Under current Australian defamation law, “innocent dissemination” is a defence. However, to access this defence, it is necessary to take down the defamatory material within a “reasonable” time after the page administrator becomes aware of it.

The bill has other impacts on the business models of social media providers. It requires any provider with a user base of more than 250,000 users to have an office or agent in Australia. This is a polite way of saying there must be someone to sue in the jurisdiction. And the bill requires social media providers to have a complaints process. If a person complains a post is defamatory, with the poster’s consent the social media firm can hand over contact details. This aspect of the bill has been criticised. As Shireen Morris wrote in the Fairfax press: “how many anonymous bullies would consent to having their identities exposed, knowing a defamation suit is imminent? If they wanted to be identified, they wouldn’t be anonymous.”

If, as seems likely, the posting troll does not consent to being exposed, the defamed person can get an “end-user information order.” This would require the social media provider to hand over contact details such as a name, a phone number, and an email address. But here’s the catch—if the contact information is not accurate, the social media provider remains liable as a “publisher” in any resulting defamation action. Those permitting comments on posts on their social media pages will no longer be liable but if the bill is enacted in its current form, the expense and risk of moderation, amplified by the risk of false contact information, will be transferred to the social media giants.

To prevent frivolous complaints, social media providers can dismiss complaints if they are not defamatory. While many commentators have equated criticism with defamation, morally and legally, they are not necessarily the same thing. Defamation is communication from one person to at least one other that harms the reputation of an identifiable third person, where the communicator (the publisher) has no legal defence. Establishing the fact of “publication” is a critical preliminary in a defamation action.

Some commentators have suggested the bill is about making it easier for politicians to bring defamation actions against their critics. However, in Roberts vs Bass the High Court held that:

Publishing material with the intention of injuring a candidate's political reputation and causing him or her to lose office is central to the electoral and democratic process. There is nothing improper about publishing relevant material with such a motive as long as the defendant is using the occasion to express his or her views about a candidate for election.

Politicians spend much of their time attempting to damage the reputations of their opponents. But qualified privilege, which covers the right to reply to damaging criticism, is an existing defence in a defamation action. The bill changes nothing in this regard.

Besides innocent dissemination and qualified privilege, other legal defences in defamation actions include honest opinion (or fair comment) and truth. Nothing in the bill reduces the opportunity for anyone to express honest opinions and fair comments to the effect that Scott Morrison, Barnaby Joyce, and every Member of the Government should be voted out of office at the next election. Nor is there anything to prevent people saying Anthony Albanese, Adam Bandt, and those in Opposition should be kept out of office at the next election. What the bill does is enable those posting defamatory comments to be exposed to legal action and held to the same standards of accountability as if they were published in traditional media. As the Prime Minister, Scott Morrison, said when he announced the bill, “The rules that exist in the real world should exist online too.”

The bill works by defining social media providers as publishers in online defamation cases but adds a new defence in law that enables them to obtain immunity in an action if they expose the identity of the person who posted the alleged defamation. The clear intent of the bill is to expose those posting comments online that are legally indefensible, the “faceless cowards” who engage in character assassination that Joyce complains about.

Reaction from the social media firms so far has been muted. After the bill was announced by the Prime Minister, Reuters sought comment from Facebook, Twitter, and YouTube. All declined. Previously, Twitter has said it routinely cooperates with legal requests for user identities, but that it values the importance of protecting whistleblowers. Facebook has said it could not reasonably be expected to monitor all comments on its website for defamation, and that it often has less access to users’ pages than the users themselves. Joyce has told them to spend some of their billions on AI to figure this out:

It is reasonable to ask that an organisation with the competency to make itself among the biggest companies in the world also has the competency to monitor the content that makes it the money. That the same companies, those putting authentic investigative journalism out of business, honour a social contract not to demean and destroy people in a way that this news masthead, if it did the same, would be taken to court for defamation.

Reaction from the opposition has also been muted so far. When the bill was announced, they said they would wait and see what was in it. They will probably do some polling before deciding what they think of it. Procedurally, the bill will now go through a review process by a Select Committee in Parliament. As a result, the bill might change. Better protections for whistleblowers would be a good idea. Parliament has finished sitting for this year, but over the Southern summer, I suspect Big Tech will be upping their lobbying budgets in Australia.

Some technical commentators have argued that trolls can get around the proposed law with a VPN. This is true. However, besides upping their lobbying budgets, Big Tech might up their AI budgets on defamation detection. It will be in their interests to start warning users making posts that their draft might be defamatory and flagging such posts for swift attention by human moderators. This is an active area of AI research.

Some lawyers have made the point that they can get information from social media providers in defamation cases already. However, their clients are typically rich enough to afford legal fees of hundreds per hour to initiate proceedings. It seems to be clear that the bill will make it much easier for ordinary people unable to afford heavy legal fees to get defamation taken down. The bottom line is that social media providers will wear the cost of the complaints mechanism and the risk of failing to reliably identify users, thereby losing immunity in defamation actions.

If enacted, this bill would significantly change the business model of social media providers. It would require them to identify users with a high degree of certainty. This remains the elephant in the room. As things stand, requiring a name, email and phone is not worth much. Names and emails can be fake and prepaid “burner” phones are easy to buy. Will users of Facebook, Twitter, and YouTube in Australia have to provide a passport, a driver’s licence, or alternative proof of address to sign up for a new account? Such onerous identification requirements to eliminate defamation would dramatically slow down new signups and wreck the exponential growth tech entrepreneurs covet. It might even motivate some social media providers to quit the Australian market entirely.

Nigel Phair, Director of the Institute of Cyber Security at UNSW Canberra said on Sky News, “It’s just not going to work, the social media companies logically don’t want this, they’ll bog it down in courts, they’re not going to want to ask their users to provide names, addresses, proper email addresses, proper phone numbers—on volume, it’s not workable.”

On the other hand, the new law might create opportunities for providers of reliable online identification to fill the breach. Firms like Facebook, Google, and LinkedIn (owned by Microsoft) already provide a degree of online identification. If other nations follow the Australian lead, they might make a virtue of necessity and find a way. A possible route might be integration with official government online ID services such as MyGovID in Australia and RealMe in New Zealand.

Some critics have noted that the bill would move Australian law closer to Chinese law. In what the historian Niall Ferguson calls “the imperial Chinese Panopticon,” pseudonymous online activity is banned completely. Speech criticising the ruling Communist Party gets people disappeared. In spite of this, the coverage of the bill by Xinhua was admirably factual, containing nothing in the way of opinion. Lately, it has been unusual for the Chinese not to put the boot into Australia. Even the Guardian, which is hardly sympathetic to the centre-right politics of Morrison and Joyce, has said that a crackdown on social media giants monetising trolling and anger is welcome.

The bill says nothing about banning pseudonyms. It would require that your social media provider knows your real name, otherwise your provider would run the risk of being liable as the publisher of any defamation you might post. To maintain privacy and pseudonymity, you would have to avoid defaming people. This does not mean you cannot express your honest opinion based on substantially true facts. You would still be free to criticise the government or whoever you liked without being exposed, provided you said nothing that crossed the line from criticism to defamation. Defamation is ruled on by the courts, not the government of the day. The proposed law will make it a lot harder for trolls to get away with malicious fabrications, venomous bullying, and character assassination online. Plenty of people from left to right on the political spectrum have suffered.

The bill is an intriguing development. It has the potential to severely disrupt the operation of many “disruptive” social media providers. Smashing the “platform” shield of social media would contribute to a restoration of traditional standards of debate in the online public square and drive bilious trolls into the depths of the dark web. The defence of daughters from “the ambit scratchings on lavatory doors” of online bullies could be a vote-winner in the parental demographics. Morrison and Joyce will campaign on the issue. With an election due by May 2022, we shall soon see if it is.

PoliticsSocial MediaFree Speech

Sean Welsh

Sean Welsh holds a PhD in philosophy. He is co-author of An Introduction to Ethics in Robotics and AI, and author of Ethics and Security Automata.