Should media have a duty of care to discourage political polarization?

This question passed my mind as I was reading Steven Pinker’s book, When Everyone Knows That Everyone Knows … Common Knowledge and the Science of Harmony, Hypocrisy and Outrage, 2025.

The question was prompted, in particular, by a statement by the Australian Prime Minister, Anthony Albanese, as reported in The Australian, at a recent UN function in New York. Mr Albanese tied his government’s proposed new digital duty of care laws to “misinformation, disinformation and increased polarisation”, saying digital services were “having an impact on our entire society and how it functions”. In subsequent media appearances, he appeared to walk away from that statement, returning to the message that the proposed laws are primarily about “giving people control” over algorithms used by social media companies.

Mr Albanese has not allayed my concerns that the proposed laws have potential to enable a future government to impose a form of political censorship. I will briefly consider some of the concerns that others have raised on that score before turning to the question of what light Pinker’s book sheds on the role of social media in political polarization and the motives for political censorship.

Proposed duty of care laws

The exposure draft of the Online Safety Amendment (Digital Duty of Care) Bill 2026 indicates that its aim is to ensure a “safe online environment”, a standard that is defined primarily in terms of content. In his submission on the exposure draft, Andrew Bushnell of the Institute of Public Affairs notes that some of the harms listed in the exposure draft require subjective interpretation. More importantly, he notes that the proposed law gives the federal government “almost limitless power to expand the range of harms covered by the safe online environment obligation”.

An indication of the kinds of speech that the federal government might regard as harmful is illustrated by the attempt to remove an X post protesting against the formation of a queer club for children as young as 8 in a Victorian school (discussed here). The author of the post argued that children of that age are too young to be exposed to gender ideology. The eSafety Commissioner sought to have the post removed by sending a complaint alert to X. In this instance, the author of the post was able to conduct a successful legal appeal. I presume that one of the aims of the proposed law is to remove the possibility of future legal appeals of that kind. The proposed law is certainly intended to give the government greater legal power to suppress controversial views.

Pinker on social media’s role in polarization

Pinker begins the relevant discussion by referring to Jonathan Haidt’s 2022 article in The Atlantic entitled, “Why the Past 10 Years of American Life Have Been Uniquely Stupid”. Haidt blamed the “virality generators” – the Like, Share, and Retweet buttons – of the social media platforms for making it easy for anonymous ignoramuses to signal their moral superiority and tribal loyalty at low cost to themselves.

Pinker builds on Haidt’s explanation by suggesting that “virality does not just exponentiate the number of viewers”, it also “creates a sense of common knowledge”. What he means by common knowledge is knowledge that A knows that B knows, and B knows that A knows and so on. That meaning of common knowledge is illustrated in the image shown above, which has been copied from Pinker’s book. It differs from the usage of the term in ordinary English to refer to facts that are widely known, irrespective of any individual’s knowledge of who knows them.

Social media platforms generate common knowledge within online communities of people who have similar interests. The common knowledge generated within those groups can become norms that hold the groups together. When members of groups express doubts about matters that have become common knowledge, they are punished by other group members – just as members of religious sects are punished when they depart from accepted dogma.

Pinker observes:

“Twenty-first-century America has become polarized into political sects, each functioning as a quasi-religious community unified by its own commonly held but not easily verified beliefs.”

The examples of polarized thinking that Pinker provides are well known. Rather than spell them out here, I ask readers to imagine a political party led by Chicken Little. Mr Little has a vast number of dedicated followers on social media who share his belief that the sky is falling. Every raindrop that falls is interpreted as confirmation that the sky is falling. Attempts to persuade party members to consider evidence that the sky is not falling are largely unsuccessful because members and their leader are bound together by their common knowledge that the sky is falling.

Mr Little could be on either the conservative or the progressive side of politics. Opponents of his views consider that the growth in number of Mr Little’s followers is a threat to social cohesion. Some argue that his views should be suppressed.

Pinker on the motives for censorship

Pinker writes:

“Freedom of speech is protected in all democracies. Yet the repression of controversial ideas is no longer a hallmark of theocracies and dictatorships but increasingly a fixture of universities, the very institutions most committed to exploring ideas.”

The veracity of the first sentence of that paragraph is dubious. For example, Australia’s constitution has no free speech guarantee – only implied freedom of political communication.

Pinker has good reasons to focus on universities because protection of freedom of speech is central to their role in the search for truth. However, much of what he writes in Chapter 8, “The Cancelling Instinct” is also relevant to the exercise of free speech in social media.

The chapter opens with a list of 17 questions raised by serious scholars, scientists and writers that have caused them to be censored, punished, fired, threatened, harassed, demonized, libelled, and in some cases physically assaulted. That list begins with the question: “Do women, on average, have a different profile of aptitudes and emotions than men?”

Pinker suggests the threat to freedom of speech emerges from two features of human psychology:

• People hold many factual beliefs not because they have been shown to be true, but because they are felt to uphold a moral order.
• The upholding of a moral order takes place in the arena of common knowledge.

In respect of the first point, laudable moral principles lurk behind the 17 questions mentioned above. The censors fail to acknowledge that the questions seek empirical information that is compatible with the moral values they wish to protect. Even if it is true, for example that groups of people differ in their average traits, it would still be unfair to discriminate against individuals for that reason.

Why are censors so prone to blur the distinction between fact and value? One of the reasons given is that people tend to engineer their bundle of beliefs towards “winning arguments” rather than “laying out logical entailments”. For example, it is easier to persuade people to oppose capital punishment by bundling the belief that it is inherently immoral with the belief that it is not an effective deterrent to murder.

Pinker suggests that despite their training in making conceptual distinctions, academics are not immune to the tendency to defend factual claims that make their moral positions more compelling. Added to this, there is the complication that the postmodern current in the humanities denies that there is any such thing as objective truth.

The threat to freedom of speech also arises because the upholding of a moral order takes place in the arena of common knowledge. Pinker writes:

“What terrifies the censors and cancelers, it appears, is not that a dangerous idea might be thought, or even expressed, but that it might become common knowledge.”

For example, a member of the speech police might not be particularly upset if they thought that they were the sole recipient of an email containing a “dangerous” idea, but be outraged if it was apparent that the idea had been communicated simultaneously to a large number of people.  Pinker notes that the metaphor, “deplatforming”, is telling “because being on a platform, visible to all, is what allows a speaker to generate common knowledge”.

Misinformation and disinformation

The cancellation and deplatforming that takes place within universities interacts with social media in interesting ways. Scientists and other scholars are often able to publicize their views on social media even though university colleagues and politicians have alleged those views to be misinformed. In some instances, the censorious actions of their university colleagues seems to have attracted additional attention to the views of controversial scholars, e.g. Bjorn Lomborg, perhaps assisting their views to become more influential.

There is, of course, a substantial amount of misinformation and disinformation published in electronic form in pursuit of objectives that are Antisemitic, Islamophobic, racist, sexist, homophobic, transphobic or designed to denigrate other community groups. It is easy to agree that promotion of such views should be condemned but it is not clear what role government should play in seeking to combat misinformation and disinformation beyond prohibition of incitement to violence and protection of children.

Media proprietors may have a moral responsibility to prevent their platforms being used to inflame polarization, but that does not justify government attempts to impose content standards. By suppressing discussion of controversial social and political issues, Australia’s proposed new duty of care laws may well impose harms on the community that far exceed the harms that these laws will prevent.