Tag Archive for: Twitter

Social media as it should be

Mathematician Cathy O’Neil once said that an algorithm is nothing more than someone’s opinion embedded in code. When we speak of the algorithms that power Facebook, X, TikTok, YouTube or Google Search, we are really talking about choices made by their owners about what information we, as users, should see. In these cases, algorithm is just a fancy name for an editorial line. Each outlet has a process of sourcing, filtering and ranking information that is structurally identical to the editorial work carried out in media—except that it is largely automated.

This automated editorial process, far more than its analogue counterpart, is concentrated in the hands of billionaires and monopolies. Moreover, it has contributed to a well-documented list of social ills, including large-scale disinformation, political polarisation and extremism, negative mental-health impacts and the defunding of journalism. Worse, social-media moguls are now doubling down, seizing the opportunity of a regulation-free operating environment under Donald Trump to roll back content-moderation programs.

But regulation alone is not enough, as Europe has discovered. If our traditional media landscape featured only a couple of outlets that each flouted the public interest, we would not think twice about using every available tool to foster media pluralism. There is no reason to accept in social media and search what we would not tolerate in legacy media.

Fortunately, alternatives are emerging. Bluesky, a younger social-media platform that recently surpassed 26 million users, was built for pluralism: anyone can create a feed based on any algorithm they choose, and anyone can subscribe to it. For users, this opens many different windows onto the world, and people can also choose their sources of content moderation to fit their preferences. Bluesky does not use your data to profile you for advertisers, and if you decide you no longer like the platform, you can move your data and followers to another provider without any disruption.

Bluesky’s potential does not stop there. The product is based on an open protocol, which means anyone can build on top of the underlying technology to create their own feeds or even entirely new social applications. While Bluesky created a Twitter-like microblogging app on this protocol, the same infrastructure can be used to run alternatives to Instagram or TikTok, or to create totally novel services—all without users having to create new accounts.

In this emerging digital world, known as the Atmosphere, so named for the underlying AT Protocol), people have begun creating social apps for everything from recipe sharing and book reviews to long-form blogging. And owing to the diversity of feeds and tools that enable communities or third parties to collaborate on content moderation, it will be much harder for harassment and disinformation campaigns to gain traction.

One can compare an open protocol to public roads and related infrastructure. They follow certain parameters but permit a great variety of creative uses. The road network can convey freight or tourists, and be used by cars, buses, or trucks. We might decide collectively to give more of it to public transportation and it generally requires only minimal adjustments to accommodate electric cars, bikes and even vehicles that had not been invented when most of it was built, such as electric scooters.

An open protocol that is operated as public infrastructure has comparable properties: our feeds are free to encompass any number of topics, reflecting any number of opinions. We can tap into social-media channels specialised for knitting, bird watching or book piles, or for more general news consumption. We can decide how our posts may or may not be used to train AI models, and we can ensure that the protocol is collectively governed, rather than being at the mercy of some billionaire’s dictatorial whims. Nobody wants to drive on a road where the fast lane is reserved for cybertrucks and the far right.

Open social media, as it is known, provides the opportunity to realise the internet’s original promise: user agency, not billionaire control. It is also a key component of national security. Many countries are now grappling with the reality that their critical digital infrastructure—social, search, commerce, advertising, browsers, operating systems and more—is subordinated to foreign, increasingly hostile, companies.

But even open protocols can become subject to corporate capture and manipulation. Bluesky itself will certainly have to contend with the usual forms of pressure from venture capitalists. As its CTO, Paul Frazee, points out, every profit-driven social-media company ‘is a future adversary’ of its own users, since it will come under pressure to prioritise profits over users’ welfare. ‘That’s why we did this whole thing, so other apps could replace us if/when it happens.’

Infrastructure may be privately provided, but it can be properly governed only by its stakeholders: openly and democratically. For this reason, we must all set our minds on building institutions that can govern a new, truly social digital infrastructure. That is why I have joined other technology and governance experts to launch the Atlas Project, a foundation whose mission is to establish open, independent social-media governance and to foster a rich ecosystem of new applications on top of the shared AT Protocol. Our goal is to become a countervailing force that can durably support social media operated in the public interest. Our launch is accompanied by the release of an open letter signed by high-profile Bluesky users such as the actor Mark Ruffalo and renowned figures in technology and academia such as Wikipedia founder Jimmy Wales and Shoshana Zuboff.

There is nothing esoteric about our digital predicaments. Despite the technology industry’s claims, social media is media, and it should be held to the same standards we expect from traditional outlets. Digital infrastructure is infrastructure, and it should be governed in the public interest.

Twitter moves against QAnon conspiracy theorists

On Wednesday, Twitter announced a series of aggressive steps it would be taking over the next week to control the rapid rise of the QAnon conspiracy movement across its platform. Twitter has already suspended 7,000 QAnon-related accounts and limited or restricted 150,000 more. It is also taking steps to prevent its own algorithms from contributing to the further spread of QAnon, including by preventing QAnon-related topics from trending, preventing QAnon-linked recommendations, and blocking QAnon-related URLs from being shared.

This is not an existential threat to QAnon, which is an explicitly political and pro-Trump conspiracy theory. As Daily Beast journalist Will Sommer has written, QAnon has established a meaningful offline presence, and is a fringe but growing force in the Republican Party.

However, Twitter’s decision to restrict—but not evict—QAnon, and expected similar actions from Facebook, could still have a major impact on the conspiracy movement’s trajectory. What happens next may determine whether the movement goes onwards and upwards into the mainstream or gets funnelled off into ever smaller and more extreme niche communities.

Twitter has explicitly linked its actions against QAnon to the risk of offline harm. In 2019, the FBI reportedly considered the movement to be a domestic terrorism threat. In recent months, the explosive growth in conspiracy theories which has accompanied the Covid-19 crisis (and which we have written about in relation to Bill Gates and ID2020 conspiracies) has also propelled QAnon to new heights—and contributed to a spate of violent incidents in the real world.

In April, a conspiracy theorist deliberately derailed a train in Los Angeles near the US Navy hospital ship Mercy in an attempt to ‘wake people up’ to what he believed was taking place on board the ship. (Precisely what he thought that was isn’t clear. USNS Mercy has been the subject of multiple QAnon-linked conspiratorial narratives, but has also been incorporated into some other conspiracy theories.) Later in April, a QAnon follower in the midst of a mental health crisis who had made threats against former vice president Joe Biden was arrested with 18 knives in her car near the USS Intrepid. This month, a Canadian QAnon follower armed with four guns including ‘assault-type’ weapons attempted to crash a pick-up truck through the front gates of Prime Minister Justin Trudeau’s official residence.

Twitter also highlighted QAnon followers’ practice of ‘swarming’ particular individuals and organisations for intense, targeted harassment. Last week, model Chrissy Teigen spoke out about QAnon’s relentless attacks on her and her partner, John Legend. The week before, it was Wayfair, a furniture company which QAnon followers decided—based on absolutely nothing—was engaging in child trafficking, precipitating days of abuse of the company and its employees. It’s telling that the Twitter spokesperson who announced the crackdown to the media did not want to be named due to concerns about becoming a target for harassment.

‘We will permanently suspend accounts Tweeting about [QAnon] topics that we know are engaged in violations of our multi-account policy, coordinating abuse around individual victims, or … attempting to evade a previous suspension—something we’ve seen more of in recent weeks’, Twitter’s safety team said in a tweet.

(Strangely, one of the most significant examples of a user evading Twitter suspensions has not yet faced any repercussions. High-profile QAnon influencer Joe M’s Twitter account, @StormisUponUs, was banned in April after accumulating more than 273,000 followers. Less than a week later he was back with an extremely thinly disguised new account, @SheepKnowMore, which has built up more than 107,000 followers in just a few months and remains active on the platform as of the time of writing.)

QAnon’s rapidly accelerating shift into the mainstream has been directly fuelled by its access to mainstream audiences on Twitter and Facebook, and by the endless rabbit warren of conspiracy content on YouTube. Its main demographic is a subset of US President Donald Trump’s base; the bulk of QAnon followers tend to be middle-aged or older, white, socially conservative and, by and large, not particularly tech-savvy.

Periodic efforts by some members of the QAnon community to encourage more users to follow them to fringe platforms such as Mewe and Gab have met with lukewarm responses. Some QAnon followers are now encouraging one another to move to Parler in response to Twitter’s announcement, but it seems probable that this will go the same way as previous attempts. The mainstream platforms are where the biggest audience and potential for influence are, and sooner or later they always drift back.

Other QAnon followers have suggested that they should turn to using ever more coded language to avoid detection. This includes the individual who poses as the ‘Q’ persona behind QAnon, who posted that followers should ‘learn the use of camouflage on the digital battlefield’.

This tactic seems unlikely to work. As the boogaloo movement is discovering on Facebook, code words are only useful if a majority of the community can understand them faster than the platforms can. A code which almost no one understands is not useful for building a mass movement, and the likelihood that most of QAnon’s mainstream converts will be able to adapt fast enough to keep ahead of Twitter’s moderators is not high.

Assuming Twitter and Facebook are serious about making their platforms an inhospitable place for QAnon to proselytise, and are willing to commit the time and resources to enforcing those policies, the likely outcome is a deceleration of but not an end to QAnon’s recent rapid growth. Influential QAnon figures and, increasingly, QAnon political candidates will keep on using the platforms to build their own followings and drum up support and funding.

Twitter’s decisions this week amount to stopping algorithmic petrol from being thrown on the fire of the QAnon conspiracy. It’s a positive step, but so long as that fire continues to burn, the embers and spot fires will continue to spread.

Snarky tweets as a national security imperative

The ‘one-tweet hacker’ slipped through the fingers of law enforcement once more last week when the Australian Federal Police dropped its investigation into the alleged hacking of Liberal candidate Jessica Whelan’s Facebook account during the federal election campaign, after she failed to provide the referral needed for the investigation to proceed.

This case is the latest in a troubling pattern of political figures attributing embarrassing social media activity to ‘hackers’, only to see law enforcement investigations either not proceed or fail to find any evidence of unauthorised access to the complainants’ accounts when they do.

When former defence minister Christopher Pyne’s and Health Minister Greg Hunt’s Twitter accounts were each caught favouriting a porn tweet, they blamed it on a hacker with the peculiar MO of breaching high-profile political accounts and using this extraordinary access to favourite a single embarrassing tweet. Australia’s ambassador to the US, Joe Hockey, did the same when his account liked a tweet criticising former PM Malcolm Turnbull. Even Prime Minister Scott Morrison claimed an ‘unauthorised’ person accessed his account when it liked a tweet questioning China’s human rights record while he was treasurer. After their initial claims, neither Morrison, Pyne nor Hockey proceeded with a formal referral to the AFP. Hunt did, but after investigation, the AFP found no evidence that his account was hacked.

I don’t call out this trend just to be snarky.

Foreign interference through cyber-enabled information operations is a real and serious threat facing Australia and other liberal democracies. The most high-profile example of this was Russia’s weaponisation of information obtained through the hacking of Hillary Clinton’s campaign and the Democratic National Committee in the lead-up to the 2016 US election. Since then, we’ve seen a flurry of similar operations around the world. ASPI’s Hacking democracies report found that 20 countries have experienced cyber-enabled foreign interference of electoral and democratic institutions since 2016. The most common form of interference wasn’t direct tampering with election results but disinformation campaigns aimed at undermining candidates, increasing polarisation and reducing public trust in institutions that underpin our democracies.

It’s entirely plausible that a hostile foreign actor could seek to embarrass a political figure by hacking their social media accounts. But when politicians cry ‘I’ve been hacked!’ in the wake of online embarrassments without giving the claim the gravity it deserves and fully engaging with law enforcement investigations, it undermines the ability of responsible actors to call out real incidents. It makes us more vulnerable to subsequent information operations by feeding ‘truth decay’—the public perception that it’s impossible to establish the objective facts of such an incident.

Preventing cyber-enabled information operations is challenging because they are cheap and easy to run and are hard to detect in a timely manner. Responding to them is difficult, too, as it’s hard to identify perpetrators with 100% certainty. The best protection we have against these attacks is healthy, robust democratic institutions that the public trusts. When functioning well, these institutions can act as a national immune system that can identify and alert the public to disinformation campaigns.

A society in which the public trusts media organisations, political parties, electoral institutions, courts, law enforcement agencies and government is the best inoculation against foreign disinformation campaigns. Trusted democratic institutions are our ‘magic weapons’.

Unfortunately, trust in democracy and politics is at a low point in Australian history. Only 41% of Australians are satisfied with the way democracy works, down from 86% in 2007. More than 60% of Australians believe the honesty and integrity of politicians are very low.

Restoring public trust and confidence in our democratic institutions should be viewed as a national security imperative for liberal democracies.

Policy reforms and institutional innovation are part of the solution to restoring this trust. The federal government should combat public perceptions of corruption by establishing a national integrity commission. We need to protect our independent media by ensuring they are free from intimidation and political interference. Our media can only be independent if they are appropriately funded. Ensuring that would mean looking seriously at the recommendations of the Australian Competition and Consumer Commission’s digital platforms inquiry and properly funding our public broadcasters.

But the personal conduct of individual actors within our democratic institutions matters too. In an environment of declining public trust, people in our democratic institutions need to constantly reflect on how their conduct affects the health of the system. When politicians make claims of hacking that aren’t validated by subsequent law enforcement investigations, their behaviour further erodes public trust in our democratic system.

When political figures face incidents of online embarrassment, we should expect them to ensure that the potential for ‘insider’ responsibility is properly investigated before they make public allegations of ‘hacking’. Similarly, it should be a strong norm that when political figures make public claims of hacking, they fully cooperate with law enforcement.

Sociologists will tell you that the best way to enforce a norm is through social sanction. The case of the ‘one-tweet hacker’ certainly attracted its share of snarky tweets—the 21st-century social sanction. Given the context, the mocking tweeters can reassure themselves that their tweets weren’t just cathartic—they were small contributions to strengthening our democratic resilience and national security.

A roadmap for reining in big tech

Anyone watching Mark Zuckerberg’s testimony to Congress in April following the Cambridge Analytica scandal would be forgiven for thinking US politicians shouldn’t be allowed anywhere near tech policy.

The questions the mostly grey-haired representatives fielded to the 33-year-old Facebook CEO ranged from the basic to the bizarre. ‘How do you sustain a business model in which users don’t pay for your service?’, asked Orrin Hatch. ‘Senator, we run ads’, was Zuckerberg’s reply.

Other gems included: ‘Why am I suddenly seeing chocolate ads all over Facebook?’, ‘Is Facebook spying on the emails I send via WhatsApp?’ and ‘Do I have as many friends as I think I do?’.

Thankfully, a policy white paper released on Monday by Democratic Senator Mark Warner, himself a former tech executive, has belatedly offered a framework for politicians to use when representatives from the big tech firms return to Washington in September.

The document lists 20 proposals compiled by Warner’s staff that they hope will ‘stir the pot and spark a wider discussion’ in the wake of the Cambridge Analytica scandal and Russian hacking of the 2016 US election, as well as a general unease about the growing power of the big tech firms.

The proposals span three major areas: dealing with the epidemic of disinformation, strengthening privacy and consumer protection, and ensuring competition in the marketplace.

Some of the ideas, including the introduction of a so-called ‘Blade Runner law’, which would require bots to be clearly and conspicuously labelled, represent technical tweaks that the platforms would likely find achievable and that some are already working towards implementing.

Other ideas call for fundamental changes to the business models of the big social media companies, and are unlikely to be met with much enthusiasm from Silicon Valley.

One such proposal involves introducing media-style rules on fairness and libel. Specifically, it suggests changing section 230 of the US Communications Decency Act to make it possible for people to sue tech platforms that don’t take down defamatory material posted by users.

This change would take the onus away from victims to search for, and report, material that defames, threatens or falsely accuses them. Instead, that responsibility would fall to the platforms, which are better placed and resourced to prevent the spread of such material. In effect, the rule change would treat Facebook as a media company, not just a platform—something it has long resisted.

Another major proposal entails introducing legislation similar to the EU’s General Data Protection Regulation (GDPR) that would significantly toughen privacy laws and prevent the use of personal data without the unambiguous and individualised consent of the user.

The proposals, which include ‘data portability, the right to be forgotten, 72-hour data breach notification, 1st party consent, and other major data protections’, would, in total, give users more transparency over and control of their own data.

Perhaps the most radical idea in the white paper is a proposal to require platforms to calculate the value of each user’s data. Such a requirement, the paper argues, would stimulate competition by providing ‘price transparency’ to consumers. It could also educate users about the true value of their data and attract new competitors with more favourable privacy provisions built into their products.

It’s easy to see how such a change might encourage users to get behind even more radical ideas that would require big tech companies to pay them to use their services and to hand over valuable data—something advocates of ‘data as labour’ have been calling for elsewhere.

The paper found that calculating the value of each user’s data could even help guide antitrust policy. Regulators could consider actions that increase the value of user data to be anticompetitive since they would be equivalent to a ‘price’ increase for users.

Another proposal would see tech platforms labelled as ‘information fiduciaries’, which means they would need to follow similar rules to legal or financial institutions.

The white paper drills down to underhanded tricks employed by tech firms known as ‘dark patterns’, which are used to corral users into accepting terms that invade their privacy and greatly benefit the service provider.

It even looks ahead at emerging problems that are likely to make fighting false information online even harder, such as ‘deepfake’ technology, which will make detecting fake news even more difficult.

It’s not hard to see why the big tech companies would baulk at many of the ideas included in the white paper—given that the major policy proposals would require a complete overhaul of their business models.

Facebook lost a million users in Europe after the GDPR was introduced. Just last week, its stock plummeted, wiping more than US$120 billion off its market capitalisation. The next day, Twitter lost 15.5% of its market value.

Part of the reason for the major market correction is that the social media platforms have been harassed into actually doing something about the harmful externalities their services are creating.

Facebook is on a media blitz to show that it’s on the front foot in dealing with Russian disinformation campaigns that use its platform, and Twitter has been culling fake accounts and bots.

The companies have calculated that some short-term pain from Wall Street is worth it if it staves off any heavy-handed regulatory action from Washington.

But how long will they continue to make that calculation before they give in to the stock market’s perverse incentives of growing engagement and user numbers at any cost?

Tech firms have been happy to turn a blind eye as long as they were making profits hand over fist. The US legislature, having so far demonstrated staggering incompetence, now has a set of ideas it can use to finally move the conversation along in a more substantive way.

The rise of #Xiplomacy and China’s strategic narrative

It’s been nearly two months since China’s official media agency Xinhua debuted #Xiplomacy on Twitter. Xinhua, with a weighty 11 million Twitter followers, initially used the hashtag to push out a new multipart documentary series on President Xi Jinping’s diplomacy (you can stream episodes 1–6 on Xinhua’s YouTube account here).

But hashtag harmony didn’t last long; by mid-September it had been co-opted, filled with tweets and memes by netizens who don’t share the #Xiplomacy vision. With Twitter banned in China—and on a social media network with an immediate and largely uncensored feedback loop—it was always going to be difficult for Xinhua to prop up the hashtag and keep it trucking in its intended direction.

Social media platforms now play an incredibly important role in how states—both overtly and covertly—engage and attempt to influence populations overseas. China has never ranked highly in the world of digital diplomacy but that’s quickly changing.

Over the past two years, the Chinese government has upped the ante with a global multimedia strategy that leverages China’s state-owned media outfits. The strategy is particularly heavy on animation, music and storytelling (and sometimes all three!)

Xinhua has developed a set of videos laying out China’s vision and point of view—on topics such as China’s South China Sea claims, its military power, the One Belt, One Road initiative and the recent 19th National Party Congress. And those videos will be promoted across the world via different platforms and multiple video-streaming sites. It’s a clever way to tell, re-tell and argue for a strategic and global narrative. And it certainly helps that many of them are fun and easy to digest.

Last month the New York Times revealed that Facebook had helped state broadcaster CCTV set up a Facebook page for President Xi Jinping. A small but very interesting detail to emerge as Facebook continues its efforts to crack the Chinese market. Called ‘Xi’s Visit’, the Facebook page provides almost daily updates on the president’s schedule, with a particular focus on his overseas trips and his attendance at multilateral events in China. It, too, turns to animated videos to explain China’s foreign policies.

The geography of President Xi Jinping’s followers also paints a particularly interesting picture. So who’s listening?

Geography of likes for ‘Xi’s Visit’ Facebook page. Graphic created by Danielle Cave, September 2017

Not Australians. And few living in western democracies. But neighbours to China’s south are, despite often low internet penetration rates, which put a third, or less, of their populations online. Xi’s friend count is highest in India (733,900), followed by Pakistan (480,000), Myanmar (356,000), the Philippines (240,000), Cambodia (161,000) and Indonesia (56,600). Not bad figures for a two-year-old Facebook account that appears to put minimal effort into recruiting new friends.

Online influence isn’t only about reach and, as Russia’s bold and sophisticated cyber interference has highlighted, it’s not only overt. Government social media accounts tend to amplify events and third-party content and they struggle with genuine engagement and influence. Building a narrative that explains who you are, what you want the world to look like and how you intend to get there is hard. Proactively and strategically communicating that narrative online in a way that’s targeted, convincing and effectual is even harder.

The Chinese government’s intensified efforts to get its narrative out there—even if not everyone is buying what it’s selling—are impressive. These efforts provide a lot of lessons for Australia as we wait for the government to release a new foreign policy white paper. While we’re unlikely to turn to rap videos to promote our foreign and defence policies, we do need to ask and answer an important set of questions, starting with: What’s Australia’s strategic narrative? Who’s listening? And how can we better cut through?

Donald Trump’s dark art of the tweet

President Donald Trump’s critics have consistently underestimated his political communication skills, perhaps because he is so different from predecessors such as Franklin D. Roosevelt and Ronald Reagan. Both FDR and Reagan, after all, were known as “great communicators.”

Although large segments of the American population hated them, FDR and Reagan addressed the American people as a whole and tried to appeal to the centre. Trump, by contrast, has appealed primarily to the minority that elected him. His inaugural address sounded like a campaign speech; and, after taking office, a series of false statements and provocative executive orders has undercut his credibility with the centre but reinforced it with his base.

Trump’s communication skills were honed in the world of reality television, where outrageous and provocative statements entertain audiences and boost viewership. He used that approach during the Republican primary to dominate attention among a crowded field of 17 candidates. By one estimate, Trump received the equivalent of $2 billion of free television advertising, swamping the $100 million in paid advertising raised by his Republican rival, Jeb Bush.

After he won the Republican nomination, many expected Trump to follow the traditional path of moving to the centre for the general election. Again, he defied expectations and focused a populist campaign on segments of the population that had lost jobs to global competition; and/or resented the cultural changes that had occurred over the past few decades. This populist appeal was effectively targeted, and he won the Electoral College, despite losing the popular vote by nearly three million votes. But for 100,000 votes in three rust belt states, he would not be president.

Given this, many observers expected him to target his messaging at the political centre after he took office. But Trump again confounded the pundits, by continuing to target his base voters. Some speculate that he is aiming to build a new populist party of working-class voters (former so-called Reagan Democrats) and Tea Party Republicans.

Trump also proved unconventional in his choice of communication instruments. New technologies open new opportunities. FDR used carefully paced public “fireside chats” made possible by radio broadcasting. Reagan was a master of the scripted speech dramatised on television. Reagan’s White House staff kept the administration’s message focused on the priority issue of the day or week. Trump used Twitter, in addition to his mastery of cable television, to leap over the heads of staff and the press and drive the public agenda during the campaign.

To the surprise of many, Trump has continued the practice in the White House. The use of Twitter was not new—Obama had a well-staffed account—but Trump’s personal involvement raised questions of how to manage policy thunderbolts from the White House and express complex policy issues (such as nuclear weapons) in 140 characters. But, as a device to communicate with his base and keep attention focused on himself, government by Twitter has allowed him to appeal over the heads of Congress and the press.

Political communication changes over time, and there are many ways to communicate effectively. The ancient Greeks had schools of rhetoric to hone their skills for the assembly. Cicero made his mark in the Roman Senate after studying oratory. Woodrow Wilson was not a gifted student as a child, but he succeeded in teaching himself oratory because he regarded it as essential for leadership. Winston Churchill often attributed his success to his mastery of the English sentence. Martin Luther King, Jr. benefited from growing up in an African-American church tradition rich in the rhythms of the spoken word.

For some it comes easier than others. Mario Cuomo, former Governor of New York, once compared Bill and Hillary Clinton: ‘She is more a Methodist, and he is more theatrical.’

Oratory and rhetoric, however, are not the only forms of effective political communication. Non-verbal signals are an important component as well. Some inspirational leaders were not great orators—witness Mahatma Gandhi. But the symbolism of Gandhi’s simple peasant clothing and lifestyle spoke louder than words. If one compares those images with pictures of the young, insecure Gandhi dressed as a proper British lawyer, one can see how carefully he understood symbolic communication.

So does Trump, in his own way. Consider his campaign’s red baseball cap with the slogan ‘Make America Great Again,’ as well as his fixation on branding when he was a businessman, and his use of Twitter.

But, in addition to communicating with distant audiences by rhetoric and symbols, leaders need the ability to communicate one on one or in small groups. In some cases, that close communication is more important than rhetoric. Organisational skill—the ability to attract and manage an effective cabinet—is difficult to combine with government by Twitter. Harry Truman was a modest orator, but compensated for the lack of public rhetoric by attracting and ably managing a stellar set of advisers.

Setting the right example is another crucial form of communication for leaders. Anticipating a sceptical public reaction when Singapore raised the salaries of government officials in 2007, Prime Minister Lee Hsien Loong announced that he would forgo the raise for himself. In terms of symbols related to conflicts of interest, Trump has not yet mastered the art of political communication.

Thus far, Trump has proven a more effective political communicator than his critics expected. But whether he can succeed in the long term with his unconventional approach is one of the great questions facing his presidency.