Keywords

Free elections do not require the absence of regulation. Indeed, regulation of the electoral process is necessary in order that it may operate effectively or at all. Not only that, but some limitations upon freedom of communication are necessary to ensure the proper working of any electoral system. (Levy v Victoria (1997) 189 CLR 579, 608 (Dawson J))

Introductory Comments

In this book we explain the need for, and then propose a practical solution to the growing problem of false election information in Australia as well as other authentic democracies. Our aim is to help clean up the ‘preference formation’ stage of elections by developing a best practice regime to manage the problem of authorised political disinformation. This, of course, excludes the vast quantity of unauthorised bad faith communication that is out there, but it is a start and a viable one. Crucially, our intent is to help clean up the election information environment in a manner that does not impair the free flow of accurate democratic speech that all democracies need in order to function properly. Neither do we want to propose any remedy that might violate the implied freedom to political communication under the Australian Constitution, not only because we do not want to recommend laws that are destined to be struck down but also because, at this point in Australian jurisprudence, we would take it as a sign that our proposed measures are probably an unjustified burden on political speech. Devising such laws is a difficult but not impossible task, as we will show.

We also offer in this book a detailed history of attempts to regulate election mis- and disinformation in Australia at both state and commonwealth levels as well as in other settings. This history is partly political history but it is also, in large part, legal history and we provide a fine-grained account of how truth in political advertising laws have been debated, enacted, implemented and understood from a legal point of view.

In this first chapter, we offer some background to the growing false election information problem in both Australia and other democratic settings and argue that it is an issue in need of urgent attention. We then do some ground-clearing preparatory to the main discussion by first clarifying the scope of our inquiry and exploring the kinds of statements that should, and can, be regulated by truth in election advertising (‘TIPA’) laws.

Background

Recent elections throughout the world have offered vivid demonstrations of the extent to which disinformation campaigns can distort democratic processes and undermine their legitimacy. The problem was especially pronounced during the 2016 US Presidential election campaign when the term ‘fake news’ entered the global political lexicon (Allcot & Gentzkow, 2017). Some of that disinformation emanated from the highest levels; for example, the Washington Post fact-checking service reported that, at the completion of his incumbency, US President, Donald Trump, had made more than 30,000 false and misleading claims during his term of office (Kessler et al., 2021). Such disinformation is thought to have changed the outcome of the election (see Gunther et al., 2016; Napoli, 2018). Democracies around the world are now witnessing the ascendance of what has been labelled ‘post-truth’ politics; a politics in which even the value of objective facts is disputed.

The media landscape and its political economy have eroded both the media’s willingness to supply, and the consumer’s demand for ‘truth’ in political discourse; fast-paced and ‘bite-sized’ communication increasingly dominates this new landscape of political communication. While social media has decreased barriers to entry and ‘democratised’ media platforms, it has also exponentially fostered the propagation of false and misleading information in the spheres of politics, and more recently, public health. This has made the consumer’s task of attempting to sift through considerable quantities of mis- and disinformation more difficult; an effect that has sometimes been exacerbated by the attitudes of those at the highest levels of government. Conspicuous examples include Kellyanne Conway’s invocation of the idea of ‘alternative facts’ (Bradner, 2017) and Rudy Giuliani’s assertion that ‘truth isn’t truth’ (Pilkington, 2018).

Repeated claims of ‘stolen elections’ during the 2020 US presidential campaign generated distrust around the integrity of the electoral process and cast widespread doubt on the authenticity of the declared result, even among certain electoral authorities (Brown, 2021). Subsequent knock-on effects included significant social conflict (most notably, the storming of the US Capitol by violent protestors in January 2021) and a generalised undermining of the authority of the electoral victor; such problems persist to the present day (December 2021). Since the 2020 US presidential election, pre-emptive and baseless accusations of electoral fraud have become increasingly normalised in other contests.1 Such a strategy is inimical to both trust in election processes and the healthy functioning of democracy.

If the election outcome is not accepted, this is a calamity in the sense that the election has failed to fulfil, arguably, its primary function: to permit the peaceful transference of power from one regime to the next. ‘Peaceful’ is the operative word here; elections are part of a procedural tradition that goes all the way back to Aristotle. According to Aristotle, by giving everyone an equal share in power in deciding who governs us, democracy operates as a mechanism for avoiding social conflict and ‘political tragedy’ (Hill, 2016). This is generally achieved by elections that are both inclusive and are perceived to have been properly run; but the outcomes of elections conducted in mis- and disinformation-rich environments are less easily accepted because voting choices based on faulty information are inauthentic. Outcomes are even less easily accepted when the false information takes the form of ‘election conspiracism’ or ‘stolen election’ type narratives. This creates instability and problems for incoming governments. Overall, then, false election information inflicts substantial damage upon the speech condition that enables democracies to function properly and elections to achieve their purposes.

These problems are not, of course, limited to the United States; in the first year following the 2016 US Presidential election, over 17 countries experienced election campaigns that were undermined by false information—‘damaging citizens’ ability to choose their leaders based on factual news and authentic debate’ (Kelly et al., 2017). During the UK 2016 Brexit referendum, the ‘Leave’ campaign misled the public by claiming that leaving the European Union would free up GBP350 million a week for healthcare—an unsupported claim that many believe tipped the public towards the ‘Leave’ option (Mason, 2017). Closer to home, the 2016 Australian Federal election was unduly influenced by false information disseminated by the Australian Labor Party (‘ALP’) in its ‘Mediscare’ campaign, which arguably eroded the Liberal Party’s parliamentary majority (Elliot & Manwaring, 2018) and affected the election result, particularly in marginal seats (Carson et al., 2019). Similarly, during the lead-up to the 2019 federal election, the Liberal-National Coalition spuriously and fatally alleged that the ALP was intending to introduce a ‘death tax’ if elected.

Misleading and inaccurate information is perpetrated by actors across the political spectrum and more recently, in highly transmissible forms through the internet and social media. Publics the world over are increasingly subject to inaccurate information, often from sources that appear to be authoritative. In Australia, over two-thirds of Australian adult news consumers have reported seeing questionable news items that they considered to be deceptive, including misleading headlines and commentary; doctored photographs and serious factual errors (Hayden & Bagga, 2018). In one 2018 Australian study, a quarter of the sample claimed that they had seen stories that were ‘completely made up’ (Hughes, 2020). These figures are corroborated by equally troubling rates of disinformation exposure in other established democracies (Loomba et al., 2021).

This dynamic is exacerbated by increasing political, economic and cultural polarisation, the distorting influence on democracy of digitisation, and the rise in power of digital platforms. Ignoring disinformation is no longer an option because the stakes have become so high. False election information can undermine the autonomy of voters, delegitimise the electoral process, alter the course of elections and undermine trust in democracy (Rowbottom, 2012). As Ari Waldman puts it: ‘democracy is under threat when the truth is no longer a check on power’ (Waldman, 2018: 869).

Although most Australians think something should be done about political disinformation, solutions are not straightforward because, while mis- and disinformation pollutes the speech environment, any attempt to limit speech risks harming the free speech condition that is vital for democracy’s functioning. Without adequate freedom of political expression, democracy dies (Kelsen, 1961). Working out what ‘adequate’ means is therefore a central concern when considering the problem of how to manage false election speech.

In this book we tackle the problem by first establishing whether it is really a problem in need of a solution. We then consider different approaches with a focus on legal solutions. After canvassing the history and character of experiments in legal responses worldwide, we then concentrate on the Australian experience. Our focus throughout is on moral justifications for a legal response as well as how such laws can, do and might operate. Of special interest is the constitutionality or otherwise of laws enacted to combat false election information.

A Home-Grown Solution

One way of combatting false and misleading election claims is by the use of ‘Truth in Political Advertising’ laws. Broadly speaking, as their name implies, TIPA laws are designed to regulate the ‘truth’ content of political advertisements or statements within political discourse. As we will show, the content and therefore functionality of these laws varies significantly throughout the world. TIPA laws exist de jure throughout many mature democracies, but arguably the most effective—or perhaps, least controversial—TIPA regime exists in South Australia. South Australia’s TIPA provisions are contained in s 113 of the Electoral Act 1985 (SA) which reads:

113—Misleading advertising

  1. (1)

    This section applies to advertisements published by any means (including radio or television).

  2. (2)

    A person who authorises, causes or permits the publication of an electoral advertisement (an advertiser) is guilty of an offence if the advertisement contains a statement purporting to be a statement of fact that is inaccurate and misleading to a material extent.

    Maximum penalty:

    1. (a)

      if the offender is a body natural person—$5000;

    2. (b)

      if the offender is a body corporate—$25,000.

  3. (3)

    However, it is a defence to a charge of an offence against subsection (2) to establish that the defendant:

    1. (a)

      took no part in determining the content of the advertisement and

    2. (b)

      could not reasonably be expected to have known that the statement to which the charge relates was inaccurate and misleading.

  4. (4)

    If the Electoral Commissioner is satisfied that an electoral advertisement contains a statement purporting to be a statement of fact that is inaccurate and misleading to a material extent, the Electoral Commissioner may request the advertiser to do one or more of the following:

    1. (a)

      withdraw the advertisement from further publication

    2. (b)

      publish a retraction in specified terms and a specified manner and form

    (and in proceedings for an offence against subsection (2) arising from the advertisement, the advertiser’s response to a request under this subsection will be taken into account in assessing any penalty to which the advertiser may be liable).

  5. (5)

    If the Supreme Court is satisfied beyond a reasonable doubt on application by the Electoral Commissioner that an electoral advertisement contains a statement purporting to be a statement of fact that is inaccurate and misleading to a material extent, the court may order the advertiser to do one or more of the following:

    1. (a)

      withdraw the advertisement from further publication;

    2. (b)

      publish a retraction in specified terms and a specified manner and form.

In this book, we offer a broad outline of an ideal-type model for TIPA laws in Australia which we base on the South Australian system because it tells us a good deal about the benefits and pitfalls of TIPA laws and what to avoid and pursue in an ideal-type regime. However, in sketching out our preferred model, we offer a number of enhancing modifications to South Australia’s framework, some of which were inspired by good practice (as well as shortcomings) in other common law jurisdictions. We conclude with a list of ten recommendations for a regime suitable for adoption in all Australian jurisdictions with a close eye on ensuring that TIPA laws do not exert a chilling effect on the kind of political speech that it is designed to protect, and which is the lifeblood of a thriving democracy. In doing so we seek to resolve the democratic conundrum, neatly captured by Joo-Cheong Tham and Keith Ewing as the ‘paradox of elections’ whereby elections require ‘both freedom and restraint if electoral purposes are to be served’ (Tham & Ewing, 2021: 312). In other words, we are attempting to find a way to deter authorised untruths in election campaigns while at the same time preserving democratic speech freedoms. This is a difficult balancing act.

But before doing so, we set out our terms of reference and the types of speech that can and should be captured under TIPA legislation. We then provide a detailed justification for regulating false election speech. This entails outlining its democratic harms; especially its implications for democratic legitimacy. We also determine the psychological, political and economic dynamics of false election information. Following this, we look at the history and state of play of legal experiments with this kind of legislation both in Australia and other common law jurisdictions and conclude with a list of principles and guidelines for a regime suitable for adoption in all Australian jurisdictions; one that will be viable and effective, acceptable to the public, likely to meet with bi- and multi-party support and, crucially, is resistant to legal—and especially constitutional—challenge.

So, first, what kinds of statements do TIPA laws seek to regulate? And how does the kind of communication it seeks to regulate map onto the scholarly understanding of common ‘information disorders’ such as ‘misinformation’ and ‘disinformation’.

Which Statements Are Captured?

Contrary to popular belief, so far TIPA laws have tended to be rather narrow in their scope. They are only applicable to authorised political advertising that is disseminated or published by a political party, person or body corporate prior to or during an election, with the potential to affect the result of the election. Captured statements can influence both the formation of opinion (the ‘preference formation’ stage of elections), as well as the expression of that opinion (the ‘preference expression’ stage of elections) in terms of the way in which the elector casts their vote. Political advertising can take a range of forms including television; radio; corflute boards or social media posts. TIPA laws do not apply to political communication that cannot be properly characterised as an ‘advertisement’. The definition of an ‘advertisement’ as a notice or public announcement does not encompass political debates, interviews or stump speeches, for example, but, as we demonstrate below, at times what looks like political commentary between elections can be deemed a form of electoral advertisement due to the likelihood that it will influence voting preferences.

Existing TIPA laws, both in Australia and around the world, are generally confined to authorised, demonstrably false statements of fact with the potential to influence the election outcome. They are not concerned with expressions of opinion or election-contingent predictions about policy outcomes. For example, in South Australia, the Supreme Court has stated that s 113 ‘is restricted to advertisements so that a person may make speeches that include statements of fact which are inaccurate and misleading’. It only applies to those who are identified as the author of a statement and does not penalise those who publish inaccurate or misleading statements of fact under an honest and reasonable mistake of fact.2 The section, in all those circumstances, is directed to a ‘very small class of persons in very narrow circumstances’ (Cameron v Becker (‘Cameron’) (1995) 120 FLR 199, 215 (Lander J)) (emphasis added).

Disinformation, Misinformation and ‘False Campaign Statements’

When discussing false campaign statements in the context of TIPA laws, it is worth attempting to map them onto the dichotomy of ‘misinformation’ and ‘disinformation’. Since the term ‘fake news’ came to the fore in the 2016 US Presidential Election, scholars and governments alike have struggled to delineate its precise attributes and distinguishing features. Use of the term ‘fake news’ has subsequently abated, and since 2018 the scholarly community has opted for the terms ‘misinformation’ and ‘disinformation’ to describe the two broadest categories of ‘information disorders’ (Wardle & Derakshan, 2017). Although the terms are often used interchangeably, they are distinct. Misinformation generally refers to the ‘inadvertent or unintentional’ spreading of false information, whereas disinformation is ‘the subset of misinformation that is deliberately propagated’ (Guess & Lyons, 2020: 11). The difference between them is fundamentally a question of intent.

Given that s 113 of the Electoral Act 1985 (SA)—a law that plays a central role in this study—is a criminal offence, one could reasonably expect it to require intention, but this is not the case. Violation of s 113 is a strict liability offence, and the available statutory and common law defences are assessed objectively. For example, the statutory defence available in s 113(3) places an evidential onus on the defendant to prove that they both:

  1. (a)

    took no part in determining the content of the advertisement and

  2. (b)

    could not reasonably be expected to have known that the statement to which the charge relates was inaccurate or misleading.

Furthermore, in Cameron, the South Australian Supreme Court affirmed that ‘there is nothing in the subject matter of s 113 of the Act which would preclude the common law defence of honest and reasonable mistake’ (Cameron: 206 (Olsson J)). Both defences are objective tests. For example, in Cameron the Supreme Court examined a purported statement of fact in an electoral advertisement by comparing it with statements made by the plaintiff in a radio interview on which the defendant claimed the advertisement was based. The court, in dismissing the appeal, concluded that ‘[the statement was] substantially at odds with the Lucas statement and simply could not have been accepted, by any reasonable person, as a fair and accurate projection of the impact of that statement’ (Cameron: 205 (Olsson J)). The function of this defence will be explored in further detail throughout the book.

To clarify, due to the strict liability condition found in s 113 (which we endorse), TIPA laws are applicable to both disinformation and misinformation (but not to malinformation, that is, information that is based in reality but whose sole purpose is to inflict harm). In terms of the way that the legislation operates, intent does not need to be proven and objective tests of reasonableness are applied to both the statutory and common law defences. It will always apply to examples of disinformation, where the defendant knew the statement was false or misleading and formulated the content of the advertisement. In a case of misinformation, where the spread of false information was inadvertent, the statutory defence would be available if the defendant did not formulate the content of the advertisement and could not have been expected to know that the statement was inaccurate or misleading according to the objective test of reasonableness. As the offence relates to both misinformation and disinformation, following Jacob Rowbottom, we will refer to this narrow class of directed statements of fact in which intent is superfluous as ‘false campaign statements’ or ‘false election information’, and when aggregated in an electoral system, as ‘false information’.

Therefore, when assessing the harms of false campaign statements, the intent of the propagating actor has little to no effect on the democratic harms imposed. A statement intended to smear a candidate will potentially affect voters’ preferences regardless of whether the author knew the statements to be false or misleading. With this in mind, we can now explore the dynamics and harms of false campaign statements (whether intentional or unintentional) and canvass their effects on elections and democracy in general.

Notes

  1. 1.

    California’s and New Jersey’s recent gubernatorial races in September and November 2021, respectively, are illustrative of the power of accusations of ‘stolen elections’. See, for example, Litt (2021).

  2. 2.

    This common law defence was first articulated in the Australian context in Proudman v Dayman (1941) 67 CLR 536.