"Fake news" was not a term many people used 18 months ago, but it is now seen as one of the greatest threats to democracy, free debate and the Western order.
As well as being a favourite term of Donald Trump, it has been named the word of the year, raised tensions between nations, and may lead to regulation of social media. And yet, nobody can agree on what it is, how much of a problem it is, and what to do about it.
The origins of fake news
Governments and powerful individuals have used information as a weapon for millennia, to boost their support and quash dissidence.
Octavian famously used a campaign of disinformation to aid his victory over Marc Anthony in the final war of the Roman Republic. In its aftermath, he changed his name to Augustus, and dispatched a flattering and youthful image of himself throughout the Empire, maintaining its use in his old age.
In the 20th century, new forms of mass communication allowed propaganda's scale and persuasive power to grow, particularly during wartime and in fascist regimes.
This sort of propaganda was largely funded and controlled by governments, but the blatant bias it carried waned as the ideological struggles became less apparent. Added to that, as populations became more used to mass communication, they could more easily see through it.
Enter the internet and social media
Before the internet, it was much more expensive to distribute information, building up trust took years, and there were much simpler definitions of what constituted news and media, making regulation or self-regulation easier.
But the rise of social media gas broken down many of the boundaries that prevented fake news from spreading in democracies, allowing anyone to disseminate information, especially those that have proven most adept at exploiting the social networks.
Facebook and Twitter allowed people to exchange information on a much greater scale than ever before, while publishing platforms like WordPress allowed anyone to create a dynamic website with ease. In short, the barriers to creating fake news were undone:
Russia, the US election and Trump
Hoaxes and falsehoods have been associated with the internet since its early days, but it is only in the last two years that organised, systematic misinformation campaigns, often linked to governments, have emerged, and their effect on democracy and society scrutinised.
Last year's US election is seen as providing a fertile breeding ground for fake news. Some credit Donald Trump's anti-establishment rhetoric and distrust of the mainstream media. Others blame widening partisanship, which meant readers were more prone to believe and share stories that fit their beliefs.
But the rise of social media itself has also been seen as central. Sites like Facebook are accused of creating "filter bubbles", the phenomenon of showing people things that they like or tend to agree with, and hiding those that they don't.
Critics of Facebook and Twitter say the sites are purpose built for spreading misinformation, with the reach of a story dependent on its ability to go viral - something that often depends on sensationalism and emotional reactions more than truth itself.
Headlines such as "Pope backs Trump", "Hillary sold weapons to ISIS", "FBI Agent Suspected in Hillary Email Leaks Found Dead" went viral on Facebook in the run up to the election, garnering thousands of shares.
But fake news goes beyond the US. According to Freedom House, 30 governments pay "opinion shapers" to promote propaganda online.
What exactly is fake news?
One of the problems of actually doing anything about fake news is that it comes in multiple variants from different actors with different motives. These include but are not limited to:
- Commercially-driven sensational content: Stories where truth is not important, but very often have absolutely no grounding in fact. Their key goal is to drive web traffic and, as a result, generate advertising income. Pop-up websites run by Macedonian teenagers fall into this category.
- Nation state-sponsored misinformation: The goal here isn't revenue, but influence. Outlets in Russia or elsewhere might produce content to swing public opinion, sow division or give the illusion of support for a particular candidate or idea, either domestically or abroad. Fabricated stories can often be mixed with true or sensationalised ones.
- Highly-partisan news sites: These can conflate fact and opinion, are nakedly supportive of one political viewpoint or party, and often position themselves as alternatives to the mainstream media.
- Social media itself: Swarms of Twitter bots posting doctored or misleading photos, adverts on Facebook paid for by Russian intelligence outfits, videos on YouTube claiming terrorist incidents are hoaxes.
- Satire or parody: Light-hearted publications such as The Onion and Daily Mash have existed well before fake news was seen as a problem.
This is not to mention news stories from mainstream outlets that turn out to be hoaxes or mistakes, but were not believed to be when published.
And of course, there's Donald Trump, who constantly refers to outlets such as CNN and the New York Times as "fake news". His administration also coined the term "alternative facts" when presented with evidence of low turnout at Trump's inauguration.
These different definitions and motives make it incredibly difficult to effectively counter fake news. Some types, such as commercially driven fake websites, have been countered by cutting off advertising, but this does not affect misinformation campaigns, for example.
Has it had an influence?
It is hard to tell. Facebook's chief executive Mark Zuckerberg initially said the idea that misinformation on Facebook influenced the election was a "crazy idea", but has backtracked, saying he regrets the comments.
The sheer scale of Facebook and Twitter - 2bn and 330m users respectively - and the hours spent on them each week suggest many eyeballs have come into contact with fake news stories or misinformation campaigns.
According to a study from Stanford University, fake news websites received 159 million visits during the month of last year's US election, and other research has shown the most widely-shared news stories during the election were fake. The majority were also pro-Trump.
There is less evidence of fake news taking off in the UK during the Brexit vote or this year's general election, although bots are believed to have been used in both.
However, there is a difference between reach and influence. The latter is a very difficult thing to measure, and this has been true of media for years - how much does the media drive beliefs, and how much does it reflect beliefs?
One less measurable effect may have been that the sheer quantity of fake news stories may have reduced trust in mainstream media - if scepticism about what people read online increases, they may not know what to think. In these situations, people tend to stick to their prejudices.
What is being done about it?
After widespread criticism that the companies are failing to deal with fake news, Twitter, Facebook and Google have all announced measures to crack down on misinformation online.
Facebook, which has faced the most scrutiny, says it is enlisting fact checkers to flag disputed stories, cutting off advertising revenue to fake news sites, and better reviewing adverts on the site. Twitter says it has become better at dealing with bots, and Google has promised better algorithms to police YouTube.
However, critics continue to argue that not enough is being done, and that the tech companies are reluctant to take action, for fear of being seen as biased, or of being seen to accept that they are publishers. Facebook is now viewed as the most powerful media platform in the world, but repeatedly says it is not a media company.
However, the matter may get taken out of their hands amid heightened political fears. Theresa May has accused Vladimir Putin's Russia of "planting fake stories" to undermine the West. Politicians in both the UK and US have launched investigations into Russian election meddling.
How do you spot it?
Spotting fake news isn't easy: a Stanford study last year found that students were shockingly bad at distinguishing between different types of material online, whether paid for, fake or legitimate.
Facebook has a useful list of ways to spot fake news, which include checking other sources and a site's URL.
News stories are often described as so good you couldn't make it up, which is often what makes them great news stories. But it's worth expressing scepticism about everything: did this really happen?