In 2018, the Guardian published an article about an Italian company that decided to cash in on Brexit. The company had developed “leave” or “stay” emojis and was set to launch an app for them. The article spoke about how British MPs labelled these emojis as “divisive and dangerous”. The catch? Everything, from the Gibraltar- based company and the writer of the piece Scherzo Primavera (or “spring joke” in Italian), was fake. Published on the 1st of April, it was one of the many April Fool’s Day hoax reports that media organizations publish every year.
For many, the 1st of April has been seen as that one day of the year that they can project their inner Onion-esque writer. In 2018, the Independent claimed that President Donald Trump had deleted his infamous Twitter account. In 2013, Australian Geographic warned tourists in New South Wales to be wary of “drop bears”- a species that looks strangely similar to koalas. As It Happens, a Canadian interview show, has an illustrious history of running hoax news stories on this particular day. In 1982, the South China Morning Post ran a story about instant water during city water shortages. One simply had to add water to a packet of unspecified powder from China to get clean drinking water. In 2016, the Harbour Times “reported” on Chief Executive C Y Leung not running for another term in office (of course, that one, ironically, did forecast the future).
While the above examples may seem harmless, there have been instances when these hoaxes have had consequences. In 2013, two Florida DJs were momentarily taken off the air after they claimed that “dihydrogen oxide” was coming out of taps in a residential area. Dihydrogen oxide is the chemical term for water. Given that not many people knew this, the report sparked panic which lead to officials having to issue statements that the water was, in fact, safe.
A Jordanian newspaper Al Ghad in 2010 reported on a UFO landing near the town of Jafr. What ensued what reminiscent of the hysteria after H. G. Wells’ War of the Worlds broadcast in 1938. The residents panicked, the mayor called for security forces to be sent in and almost evacuated the 13,000 inhabitants of the town. In 2003, a 14-year-old was arrested for fabricating a report about the SARS outbreak in Hong Kong. Using the Ming Pao news logo, he created a fake website that had details of Chief Executive Tung Chee-hwa’s resignation, the collapse of the HSI and how the city had become quarantined.
In 2016, China banned April Fool’s pranks from social media as a whole, stating that they led to the spread of harmful rumours. This year, even Microsoft sent out a company-wide memo reminding people that pranks or “April Fool’s Day stunts” are not encouraged by the management. The memo stated, “Sometimes the outcomes are amusing and sometimes they’re not”.
With certain political leaders decrying the spread of misinformation, one would think that perhaps media outlets too should take a page out of Microsoft’s book- lest a member of their staff go all out with accounts of mountains in Boston spewing lava.
However, according to researchers Edward Dearden and Alistair Baron from Lancaster University, these hoax reports may actually help with the identification of malicious fake news stories. Their research paper, titled “Fool’s Errand: Looking at April Fools Hoaxes as Disinformation through the Lens of Deception and Humour”, highlights how the similarities between hoax stories and fake news can help curtail the spread of misinformation.
A press release revealed that Dearden and Baron compiled a dataset of more 500 articles published over the past 14 years across 370 websites. They compiled a list of common charcteristics between hoax reports and fake news stories that included the use of personal pronouns, the general length of the pieces and the use of formal language.
They were able to create a “machine learning classifier” that draws on stylistic similarities between the two genres of articles. It can classify writing as either a hoax story, fake news or actual reportage. According to the release, the classifier could identify fake news stories with an accuracy of 72 % and hoax stories with a 75 % accuracy. The paper will be presented at the 20th International Conference on Computational Linguistics and Intelligent Text Processing this April in France.