In 1934, the world’s largest plane took to the skies of Moscow. The Tupolev ANT-20, better known as the Maxim Gorky, boasted a unique fit-out. Equipped with a speaker to broadcast to thousands of outlying villages (dubbed the ‘Voice from the Heavens’), the Maxim Gorky also carried a full printing press, a complete photographic laboratory, equipment for projecting films, and according to at least one report the ‘sensational device’ of a ‘“Times Square-type” display attached to the underside of the wing for disseminating Communist slogans with red light bulbs’. It was, in short, a propaganda machine, pushing out Joseph Stalin’s latest dictates at scale. In an era when the world’s richest man is CEO of the world’s largest rocket company and owns a major social media platform, and has firm and eccentric views of his own, does the Maxim Gorky fly again?
It is in this context that the Australian government has tabled the Communications Legislation Amendment (Combatting Misinformation and Disinformation) Bill 2023. The Bill does a variety of useful things, such as requiring social media companies to report to regulators and take steps to improve media literacy. But it also creates a system of codes for dealing with the problems of mis- and disinformation, and financial consequences for companies that fail to comply. It’s hardly a Ministry of Truth situation, but it’s also not nothing. Content moderation systems at scale routinely capture unintended examples, and in times of tense political debate such models risk exacerbating division and disquiet. The Bill is a mess, both overreaching and underpowered, with spurious distinctions and exceptions making visible the contradictory conditions of contemporary public life.
Political leaders both large and small have always sought to control information flows, whether to persuade or coerce. Not all these efforts have been as blunt as Stalin’s. In liberal democracies they have often involved, to varying degrees, negotiations with the free press. As gatekeepers of information flows in the twentieth century, broadcast and publishing media have at times played a contradictory role, with these companies and the journalists working within them both courting power and holding it to account.
The problems of mis- and disinformation are a symptom of the decomposition of these power relations. The accepted definition of mis- and disinformation covers digital content that is verifiably false or misleading or deceptive, shared on digital platforms, that is likely to cause harm (the distinction between the two arising from intention). There is a right-wing talking point that misinformation is anything the government disagrees with, which is not entirely true, but not easily falsifiable. Misleading information at scale, after all, has been around since the printing press. Politicians of all persuasions lie, and yet only some of it is labelled disinformation. The debate about what constitutes mis- and disinformation ends up itself being a propaganda point—a jostle for who gets to set the frame of the Overton window in a world where the marginal cost of reproducing content is zero.
In this context, it is important to acknowledge the revitalisation of small publishers and independent journalists and analysts who have exploited accessible digital publishing capacity to impressive effect. However, it is also fair to say that the information ecosystem has benefited less than we might have hoped from the libertarian dreams of the early internet evangelists. When John Perry Barlow proclaimed in 1996 that cyberspace was independent of antiquated conceptions of government—those ‘giants of flesh and steel’—he spoke of a ‘global social space we are building to be naturally independent of the tyrannies you seek to impose on us’. Governments may be struggling to shape information dissemination in cyberspace, yet we are nonetheless dependent on tyrannies of a distinctly twenty-first-century kind.
It is no longer the privilege of the newspaper editor to shape public opinion, but of the techbro tweaking algorithms for profit. In this sense, mis- and disinformation are distinct from propaganda because the conditions that give rise to them are less direct; they are a secondary consequence of a specific business model. Digital platforms have a strategic interest in optimising features designed to attract new users, retain attention and increase interaction with the platform. The primary metric for success is engagement, because it is good for advertisers. Platforms offer and use micro-targeting, automated searches, active curation and algorithmic recommendation systems to amplify the most engaging content regardless of its truthfulness or users’ motivations for engaging with it. They do this because it is profitable.
Revenue sharing systems, in turn, create direct financial incentives for content creators to create and share engaging material. This is why, in the wake of the Bondi Junction massacre, Islamophobes, white supremacists, pro-Russia figures, Zionists and British personalities flooded social media with conspiracy theories about the identity of the attacker. Researchers in the United States have identified the twelve people primarily responsible for the vast majority of anti-vaccine conspiracies online, and found that many of them had financial incentives to continue sharing such content. Content creators are guided by the same motivations as the platforms they use.
The algorithmic amplification of such content is exacerbated by commercial business media models beyond social media that report on it in service of their own ambitions for engagement and ratings. Platforms have also become the de facto distribution system for all forms of media content, including public service media, and this ends up shaping how journalistic content is drafted. It is for this reason that the most viewed link on Facebook in early 2021 was an article that implied, without any evidence, that the COVID-19 vaccine may have been responsible for a Florida doctor’s death. The article was not from a conspiracy site, but rather the reputable masthead the Chicago Tribune. Mis- and disinformation are not simply social media problems; they reflect a dynamic between old and new forms of media which is at once antagonistic and mutually reinforcing.
Authoritarians and state-based actors have been quick to capitalise on the self-absorbed liberalism and base profit motives of social media platforms. Facebook has been credibly accused of contributing to genocide in Myanmar by allowing misinformation to proliferate on the platform. Without even the basic capacity to moderate content, the platform was described as a ‘useful instrument’ by a UN fact-finding mission for the role it played in justifying the terrorisation and displacement of hundreds of thousands of Rohingya. The company eventually admitted it had not done enough to stop this. Meanwhile, Russia has entire agencies churning out disinformation campaigns. It makes the Maxim Gorky look like a Myspace page.
The result is that for all practical purposes, the media ecosystem is increasingly being shaped by a commercial model that profits from the amplification of polarising, sensationalist content regardless of its accuracy. This is not a model compatible with democracy’s need for well-informed citizens and meaningful political deliberation. It is in direct opposition to the objective of class-conscious social movements, which is to draw out connections across tradition social divides. Consideration of collective concerns such as the public interest, human rights and community obligations cannot compete with these intrinsic, financial motivations, and in practice are not prioritised by platforms.
Perhaps, therefore, instead of Barlow’s manifesto, the better cultural artefact for understanding the moment is the Californian Ideology, in which Richard Barbrook and Andy Cameron mapped out Silicon Valley’s unlikely political heritage, made up of the ‘disciplines of market economics and the freedoms of hippie artisanship’. The central tenet of the Californian Ideology is that the web as many of us use it today was built by an elite possessed of a ‘contradictory mix of technological determinism and libertarian individualism’. The digital spaces where publics gather are run by private companies with an interest in corralling us into filtered bubbles and encouraging us to relax into spheres of soothing content. Mis- and disinformation are a function of technological decadence and a smug assumption about the universality of American values.
Many attempts to deal with misleading, harmful content shared at scale fail to grapple with this political economy of the web. Instead, the response to mis- and disinformation is a matter of whack-a-mole, which at best takes the form of considered explainers and at worst overly-officious processes of fact checking. Because while there is plenty of mis-/disinformation that falls squarely into the camp of ‘verifiably false’, it is also fair to say that the definition has at times been interpreted subjectively. Vaccines are the obvious example, with plenty of clear-cut conspiracy theory content alongside relatively reasonable speculation about vaccine risk and injury which give rise to a significant risk of harm when shared at scale. Sanctimonious libertarians and ultra-leftists claim that attempts to moderate such debates are authoritarian cosplay—that we have to counter bad speech with more speech, and resist the censorship compulsion. It is a tedious position justifying inaction, made to look credible because so many of the policy proposals of the mainstream political class lack coherence or effectiveness (and sometimes both). Too many policies in respect of mis- and disinformation look suspiciously like censorship, and while the framing is understandable, it doesn’t make it any less dangerous.
Australia’s bill is at once too much and too little. The regulatory impulse is not the problem; the application is. The proposal would give the Australian Communications and Media Authority serious powers without a proper requirement to consider human rights, and without sufficient transparency or reporting obligations. And yet these efforts only target the symptoms of the problem rather than the cause. Specifically, the definition of serious harm is overly broad, and includes harm to the Australian economy. Tellingly, the proposal also excludes professional news content entirely. There exist very real risks of both overreach and a lack of ambition in pursuit of an over-simplified solution to the problems of mis- and disinformation.
What is to be done? The most useful and courageous regulatory intervention would target the source of the problem—the commercial exploitation of personal information. At present, Australia’s privacy laws have not been meaningfully updated in four decades. Strict controls on data collection and use would force social media companies to reorganise their business model away from engagement at all costs. Such reforms would be both rights-respecting and immensely popular. Privacy reform will not solve all the problems created by social media platforms, but it represents a significant and necessary first step. The problem is that an array of extremely powerful media companies, both social and traditional, as well as advertising agencies and the entire data analytics industry have an interest in opposing such reform. We must organise and resist.
Like one of Elon Musk’s rockets, the Maxim Gorky ended up crashing in a ball of flames after an exhibition flight went wrong. It remains unclear if it was an accident or an act of sabotage. Despite repeated vows to do so, another plane like it was never built. As one historian described it, ‘it eventually faded away into obscurity as a childless prototype’. Even the most formidable propaganda machines are fallible, and it is within our grasp to make the predatory business models of social media childless prototypes.