Social Media
A Blessing and a Curse

Freedom of expression is an essential pillar of democracy. But what if the most important source of information and space for free speech – the internet – is used to pursue anti-democratic goals?
 

By Dan Gillmor

Are social media a boon to democracy or a menace? Are traditional media a boon or a menace? Is content creation by artificial intelligence (AI) a boon or menace?

The answer to all of those questions is the same: yes.

Here’s a better question: are the people who live in the world’s democracies willing to recognise the infinite nuances embedded in the widening debate over how “the media” affects our societies?

The answer to that one isn’t clear yet. So far, however, our collective answer has been – at best: maybe.

To grasp how media of various kinds are shaping liberal democracy – including citizens’ rights and responsibilities, and how they participate – we need first to understand how media have changed in the past several decades. The shift from analogue to digital has shaped public discourse and knowledge, with countless positive outcomes but – as we have seen so starkly with the explosion of politically minded misinformation – profoundly negative ones as well.

Democratisation of Creation

The internet has subsumed much of what we used to call print and broadcast media. Publications, radio, television, film, and other forms still exist, of course, but their creation and increasingly their distribution has become digital. Think of 20th century media as a manufacturing model, in which people create things and distribute them to consumers.

The 21st century media model is radically different. It incorporates the manufacturing and distribution system and expands on it in ways that could not have existed before the internet, at least not to the degree they do today. The essential element in the internet’s rise was its decentralisation of control. Media were no longer a collection of one-to-many enterprises; they evolved to become many-to-many (and many-to-few, few-to-few, few-to-many, and so on) systems. In a crucial but under-appreciated way, the distribution part of the old model shifted: now we were creating media, putting it somewhere online, and urging people to come and get it.

The word “creating” is key: We have all become creators of media. We can tell potentially huge audiences what we know. The tools for creating media are thoroughly democratised, and ever more powerful. Even the simple act of sharing things via social media is also an act of creation, with consequences we are only just beginning to understand.

The democratisation of creation and access has led to many wonderful things, not least people’s ability to share vital information in small communities (of geography and interest). As traditional journalism has shrunk at the local level, Facebook groups and other localised services have become essential places in which people can tell each other what is happening. Some of those online gatherings devolve into bickering, but they – and so much else online (Wikipedia being a prime example) – have given us easy-to-access insights and knowledge we never had before the Internet.

Erosion of Trust in All Information

As has been the case with all tools throughout history, it’s not just people of good will who use them. The refined and powerful tools of media creation and distribution are now part of the arsenals of dangerously capable people too.

The bad actors quickly latched onto social media – an increasingly key way by which people consume information and understand the world around them. Wherever you go, people are staring down at their mobile phones, often getting the “latest” from social media sites. Much of what they get via those sites is from traditional media, of course, but how it arrives on the screen is subject to manipulation by social media companies, via programming algorithms and people who use platforms to spread information, sometimes with malign intent.

The malign actors use these increasingly agile tools to erode people’s trust in all information – other than what demagogues tell their cult followers, of course. They are ideologues and saboteurs, sometimes state-sponsored. They also have free speech rights, at least in the United States. And they count on the dominant platforms’ understandable hesitancy to remove what people post there – given that their business models thrive on engagement.

But as those platforms have grown to unprecedented size, they’ve become core players in an increasingly re-centralised media ecosystem. So we have to ask: do we want Meta (owner of Facebook, Instagram, WhatsApp, and more), Google (YouTube), Twitter, TikTok, and other giant companies, to be the editors of the internet? That’s what more and more people are demanding.

Social media companies have responded to those demands in most ways by removing what they consider to be extremist or dangerous content. That is their right, and by many accounts, their duty. They have also taken down manifestly important media content, such as video evidence of human rights violations posted by activists and victims. Social media critics are asking for something that can’t be done – it is impossible to moderate user-posted content on a wide scale – and, for the most part, seem comfortable with the collateral damage that results. Meanwhile, it’s still relatively easy to find online videos of vicious liars who want to poison public debate, not to mention grisly violence such as beheadings, because the malicious actors have other options too.

Propaganda and lies to influence elections have been part of the political landscape as long as we’ve had media and voting. But digital technology gives the bad actors new weapons. Recent document leaks to journalists revealed the existence, for example, of the now infamous “Team Jorge” that bragged about its meddling in a number of recent elections, using sophisticated digital disinformation tactics.

Highly Useful and Downright Horrific

If all this wasn’t enough, the advent of “generative artificial intelligence (AI)” – so far best known from the sudden emergence of tools like ChatGPT, Bing AI, and image-creator DALL-E – has become a new shock to the system. The chat-AI text creators are poorly understood among the general public. They cannot think or reason; what they do, with sometimes alarmingly fluent style, is to predict the word most likely to follow the preceding one in a sentence. They predict this based on massive databases of information vacuumed up from the internet and other sources to form “large language models”, which are used by AI companies to train their systems and media-creation tools.

Alarmingly, these tools produce lies as fluently and persuasively as truth, which makes them a combination of entertaining, highly useful, and downright horrific. In the latter category are plausible scenarios where the bad actors use them to generate more and more persuasive lies, tailored to individual people.

Amid the raging debates over how misinformation and malicious content may be eroding the longstanding norms of liberal democracies and liberalism itself, democracies have paid far too little attention to one of the most malign actors of all: traditional media companies. In particular, though far from being the only malefactor, the Murdoch family’s media holdings have done enormous damage. Documents from a defamation lawsuit in the U.S. have revealed with stark clarity what anyone paying attention already knew: Fox “News” has deliberately injected poison into public discourse in order to accumulate power and money.

What can democracies do about all of this? Imposing draconian censorship on media – including individual people who have important information to share – could improve the supply of information, but at the cost of a basic liberty, freedom of expression.

The better answer is to address demand, to improve people’s ability to handle the information deluge in ways that give them better tools to spot and weed out misinformation and sleazy propaganda. The demand side of the supply-and-demand equation is media literacy, and we don’t provide nearly enough of it in most democracies.