On 2 February 2024 I wrote a post entitled “Truth, post-truth and post-fake”. I started the post like this: “I was sitting at my desk trying to think about something I could blog about. For some reason the word ‘truth’ popped into my head. After that I … messed about on the news database Nexis, rummaged in the Oxford English Dictionary and looked at Wikipedia, all the while reading tweets on the side.”
More than two years on, I was yet again sitting at my desk trying to think about something I could blog about…. This time the word ‘fake’ popped into my head, having completely forgotten my previous post. Instead of looking at tweets, I looked at posts on Bluesky and saw that the editor in chief of The Guardian, Katherine Viner, had written an article entitled “How to survive the information crisis: ‘We once talked about fake news – now reality itself feels fake’”. That felt true and I suddenly remembered my old post. I began to wonder what has changed between 2024 and 2026 and how can the phrase ‘post-truth’ feel so passé now, while thinking about a ‘fake reality’ feels so true?
In my old post I focused mainly on the phrase ‘post-truth’ but also talked briefly about ‘post-fake’ world. In this post, I’ll focus on the word ‘fake’ in particular. I’ll do a bit of digging into its past and I’ll speculate a bit about its future. I’ll begin, as always with rummaging in the Oxford English Dictionary, followed by a quick look some scholarly articles and exploring the word on the not so fake news database Nexis.
Origins
The etymological origins of the word ‘fake’, whether noun, adjective or verb, are rather murky and rather mythical. The OED first attests ‘fake’ as an adjective meaning “spurious” or “counterfeit” in a 1775 letter. By 1819 it was being used as thieves’ slang for “doing something for the purpose of deception”. Linguist Anatoly Liberman, writing for Oxford University Press, argues that the word’s best-traceable origin is, indeed, London criminal underworld slang, although how it got there is unclear.
Surprisingly, the phrase “fake news”, meaning deliberately misleading journalism, is attested as far back as the 1890s, long before its recent surge in popularity. Let’s look at ‘fake news’, first more generally, then in the news.
From false news to fake news
Before ‘fake news’ entered wide-spread use, the dominant phrase for this phenomenon was ‘false news’, traceable to the 16th century. However, three hundred years later, and in particular after 2016, that is, after Brexit and the election of Donald Trump to the US presidency, ‘fake news’ became a real buzzword.
PolitiFact selected “fake news” as their ‘Lie of the Year’, Macquarie Dictionary’s named it word of the year, and Oxford Dictionaries chose “post-truth” as its word of the year, reflecting how the whole semantic cluster around falseness, fakeness, fakery, truth and reality was shifting.
Before 2016, “fake news” was a relatively technical media-literacy term used to describe specific Russian disinfo operations or clickbait sites. Then Trump weaponised it against legitimate journalism, using the term ‘fake news media’, and the term essentially destroyed itself. It became unfalsifiable, a way of dismissing anything inconvenient, especially inconvenient truth. It was almost as if the English language was cracking under pressure.
This process accelerated after the advent of generative AI in the early 2020s after which the border between what is ‘real’, especially online, and what is digitally fabricated or AI generated became increasingly porous. It’s becoming increasingly difficult, for example, to distinguish between real and AI-generated images.
Since then fake news, together with alternative facts, misinformation and disinformation, has become a huge topic for scholarly discussion, from tracing its history to discussing it in the context of psychology – see for example the work by Sander van der Linden and the Cambridge fake news and misinformation research group.
Closer to home, there is now also a discussion emerging about fake articles, fake references and even fake DOIs in academic publishing, a discussion that maps neatly onto both the ethics of AI and the ethics of scholarly research in the context of corporate economic publishing power. There are direct links to what’s happening with AI slop. And just while I was writing this, it turns out that an author of book on the future of truth in the era of AI used AI to help him write the book and AI fabricated and misattributed many quotes in the book….
In 2018, Johan Farkas and Jannick Schou explored the phrase ‘fake news’ as a ‘floating signifier’ whose meaning is contested and constantly weaponised by both political left and right to discredit opponents in a fight over who (politicians, journalists, citizens) has the power to classify what is s true and false. In this context, we must constantly ask ourselves not just “is that story, image, video, article, reference etc. fake”, but “how would we even know?” Knowledge, facts and reality are threatened by encroaching fakes and, in fact, deepfakes.
Before we come to deepfakes, let us first follow some traces of ‘fake’ and ‘fake news’ in the news.
Fake news in the news
To see how phrases like ‘fake news’ were used in the news, I consulted the news database Nexis and first searched just for the word ‘fake’ in all available news (on 13 May 2026). On this database ‘fake’ first made its appearance on 13 October 1936 as recorded in “Primary Sources in U.S. Presidential History” in the sentence “safeguarding innocent investors against fake securities” – basically scams to steal money from investors. Here the link to criminal slang is still visible.
The word gained in popularity over time, as the following graph shows which plots its use in the news from 1 January 1936 to 31 December 2025. As one can see, the word gained rapid popularity between 2010 and 2020, with a steep increase in use after around 2016.
As for ‘fake news’, the phrase first seems to have been used in the 1970s, as in “’You can always fill up a program with speeches and film from hearings in Washington,’ he says. ‘But that’s a kind of fake news, unless you explain to the audience how the speech or hearing affects their lives.” (The Associated Press, 17 April 1979) The argument is that unlike ‘true news’, ‘fake news’ doesn’t provide context and doesn’t establish relevance for readers. It’s lazy news.
The phrase ‘fake news’ as it’s known today only came into its own between 2015 (when it was used in 1343 articles) and 2019 (when it appeared in 132918 articles)! As we have seen, this was largely due to the rise of Trump and Brexit. Things changed again with the advent of AI and deepfakes.
From fakes to deepfakes
Our current linguistic, social, political and epistemic crisis was exacerbated by the advent of deepfakes, which appeared on the horizon in around 2018.
Deepfakes are defined by the OED as “any of various media, esp. a video, that has been digitally manipulated to replace one person’s likeness convincingly with that of another, often used maliciously to show someone doing something that he or she did not do.” In a sense, deepfakes are digital forgeries. The first example provided by the OED is from 2018 and refers to “fake videos”. A 2022 conference paper describes deepfakes “as realistic videos with ‘false content’” (and here is a great overview of deepfakes on CBBC!).
Again, like ‘fake news’, ‘deepfakes’ are now a major object of study in the computational and social sciences, as well as humanities. One article from the field of semiotics argues that “In the digital world, human culture enters the domain of the ‘absolute fake’”. Amongst other things, this means that “everything that can be the object of digital representations, can also be the object of it without any ontological reference”.
Reality gradually dissolves and we reach a state called by some “the ontological quandary of deepfakes”. This refers to the epistemic and even metaphysical consequences “of technologies that synthesize increasingly unbounded digital unrealities”. Unrealities! What traces of this ontological quandary can we find in the news?
Deepfakes in the news
The term ‘deepfake’ is much younger than that of ‘fake news’ and was first used in the news in 2018 after which it literally shot to prominence. It reached its zenith in 2024 before losing some of its appeal.
“Deepfake” in the news, 1 January 1970 to 31 December 2025 based on Nexis data
One study of deepfakes, confirming this explosive trend, showed that “the number of fake videos online had grown exponentially since 2018, doubling approximately every six months, with 85,047 deepfake videos detected by Sensity in December 2020”.
Initially, many of the AI generated images and videos involved ‘swapping faces’ and, interestingly, the Chinese word for deepfakes, “huanlian”, translates as “changing or swapping faces”. The word contains no equivalent of ‘fake’ which researchers argue partly explains why Chinese regulatory responses focused on practical fraud risks rather than epistemic anxieties about truth. Such anxieties are, by contrast, all-pervasive in Western countries.
Deepfakes and the fake-reality turn
With deepfakes we seem to cross over from a political and social crisis to a deep epistemic and ontological one. How do we know and what do we know about what? In October 2025 UNESCO published a report entitled “Deepfakes and the crisis of knowing”, which argues that as “deepfakes blur reality, education must go beyond detection, teaching students to navigate truth, knowledge, and AI-mediated uncertainty”.
This becomes ever more difficult in a world of what one writer calls “deepfakism” — something broader than deepfakes as a technical phenomenon, amounting to an ontological crisis in how reality itself is understood and inhabited.
Let’s come back to the Guardian article that set me on the trail of ‘fake’. This is what Katherine Viner said about deepfakism without using the word: “As tech companies have prioritised capturing attention, truth has been downgraded. AI slop and deepfakes are now so rampant that it feels that your brain can no longer compute what it’s seeing. You start to question things that turn out to be true. It doesn’t help that reality itself has become so much stranger and more grotesque.” In such a context finding a ‘shared reality’ is increasingly difficult. And without that disinformation and division can spread like wildfire and mutual trust evaporates.
From etymology to ontology
As we noted at the beginning, the word ‘fake’ has rather murky, unverifiable etymological origins. The word might have originated in thieves’ slang, but even that is not clear. In a way, ‘fake’ has always been a little bit fake itself. Now, the word ‘fake’, in combination with ‘news’ and ‘deep’ has become a word that rather aptly describes an era of murky, unverifiable reality and ontology. What about the future of this word and this world? We don’t really know yet where the word and the world are taking us – the future is as murky as fake’s origins.
This post has, for once, not been about metaphor, but thinking about the implications of a fake world made me think about a famous saying: “The price of metaphor is eternal vigilance”. This means that we must constantly scrutinise the metaphors we use, as they can obscure, distort, or simplify complex realities. In the brave new world of fake news and deep fakes ‘eternal vigilance’ is called for not only with relation to metaphors but with relation to any sort of text, image and video. Reality has always been complex, but now, with the advent of generative AI, it’s even more so and we must be even more vigilant.
PS Just after I had drafted this post I read Eva Wiseman’s column in The Guardian weekend magazine entitled “A new era of fakery, where nothing is real but nobody minds” – this is a rather different take on the ‘fakery’ issue. She writes: “The thing about fakery is, we accept it because it affirms some truth we already believe”… Discuss…
Image: Photo of a trompe l’œil of a llama on a barn by Batista helder – Wikimedia Commons
Making Science Public 
Leave a comment