Digital Publishing
6 mins read

A century after fact checking’s humble (female) beginnings, it may just save journalism

In 2024, the world will go to the ballot box. From the USA to the UK, India, Russia, Indonesia, Taiwan, Ukraine and South Africa will hold elections. But how will voters know if what they see and hear is real? False information is already a serious risk; now generative AI can help bad actors suppress votes, libel candidates, create bogus narratives and incite violence. Can fact-checkers change the role of journalism by countering false information?

Perfectly created deepfakes are, perhaps surprisingly, not (yet) playing as big a role in the spreading of false information as expected. But this is nothing to celebrate.

It’s not because it is impossible to create nearly perfect deep fakes, but because they are not necessary, says Florian Schmidt, verification officer and manager of the Austria Press Agency’s fact-checking unit.

“People believe fakes even if they are cheaply produced; even if they are fakes that got produced in five minutes. And even if I can debunk them in five minutes or two minutes, people won’t believe that because they want to prove their own perspective of the world,” Florian says.

“Of course, there is a big danger when it comes to the recent developments in artificial intelligence, because it’s getting easier to produce AI-generated images or videos that really look perfect. And so my neighbour with no skills in photoshop or editing could type one sentence in DALL-E or Midjourney and get really perfect pictures.”

The explosion of generative AI tools onto the technology scene in the last year or so has catapulted fact-checkers into an ever-faster race with misinformation spreaders.

“The technology gets more and more perfect and we have to look for other approaches to debunk images and videos,” Florian says. “And we as fact-checkers are always one step behind the creators of fakes.”

At least 70 elections are taking place place across the world in 2024, shaping global politics and involving about half of the world’s adults. In Europe, nine parliamentary elections will take place, of which four will likely result in a notable change. That is not to mention India, the USA, Taiwan, South Africa, Russia, Ukraine and Iran. It is the world’s biggest election year yet – and the first time most countries will have elections since powerful generative AI tools exploded onto the scene.

Austria, too, has crucial elections in 2024. The populist, far-right Freedom Party could become the country’s strongest party in the national elections set for the autumn.

He is very concerned about the role AI will play in all these elections, Florian says. “I think AI will play a big role. The technology is there. The technology has made huge improvements since last year and two years ago.

“And I think there will be a lot of pictures and maybe even videos where you can see something, where you can see politicians say something or maybe audio where politicians argue something and this can be a big problem.”

Despite the evolving dangers, the enormity of the fact-checking task and conspiracy bubbles, Florian is remarkably upbeat about the future of journalism. Artificial intelligence and the onslaught of mis- and disinformation, he believes, may offer a big chance for journalists to regain trust.

Women first

It was only just over a century ago – long before artificial intelligence or deepfakes or indeed online news and social media platforms were conceivable – that the journalism industry in American started to focus on facts. In a fascinating piece on the history of fact-checking, the Columbia Journalism Review pointed out that early newspaper printers had more interest in opinion and polemic than objectivity. “There was little premium on facts – readers wanted the news, but they wanted it slanted.”

Wire services – the Associated Press in particular – changed their focus on getting out facts in as few words as possible (and with no opinion). Accuracy started to matter. But it was most likely the fledgling Time magazine that first hired people specifically to check for accuracy before publishing. Time, which was started in 1923, wrote in a 2017 account of the origins of fact-checking that perhaps the earliest use of the phrase “fact-checker” can be found in an ad for Time in 1938 that mentioned “the expansion of ‘its researchers and fact-checkers from ten to twenty-two”.

These first fact-checkers were all women – remarkable, given they were challenging the then all-male editors and writers. (Fact-checking remained women’s work through to the 1970s.) In the early days they used the New York Public Library as their main source of information, lugging newspaper clippings, a dictionary, thesaurus and Bible as well as a copy of Who’s Who and the World Almanac to the printer on deadline.

The New Yorker also started vigorously checking facts after a rather disastrous 1927 profile riddled with errors and the publication established perhaps the most famous fact-checking department in publishing history.

The job has changed considerably in the last century, with political fact-checking surging since the 1980s. And publishing online, and then social media platforms, in turn sparked the growth of dedicated fact-checking sites.

Florian Schmidt, verification officer at the Austria Press Agency.

Tsunami of trash

When the Covid-19 pandemic swept the world early in 2020, a veritable tsunami of false information about the coronavirus followed suit and the role of fact-checking organisations became increasingly important.

The Austria Press Agency or Austria Presse Agentur (APA), too, started its own fact-checking unit in 2020 when it saw false information, especially on a technical level, increase alarmingly and become more complex, Florian says.

“Before the pandemic, misinformation was just there. It wasn’t such a big problem, it wasn’t a big issue. But with the pandemic, we recognised that there was a huge increase in false information because it was a very problematic situation for every one of us and the whole population.

“So the people were searching for answers and the science couldn’t give them answers and politics couldn’t give them answers. So they tended to [believe] narratives that were fake answers and not scientific answers. And this is something why false information got spread so much in the pandemic.”

This trend has, of course, snowballed with the war in Ukraine and Israel-Gaza war.

“It’s very sad,” Florian says, “but I think it’s something that doesn’t go away because people are searching for those narratives to support their own perspective of the world.”

On why anyone would even come up with the idea to creative deepfakes or AI images, Florian says there are different reasons. It could be that someone wants to push a political agenda to impact election results, to profit in some way or another or “just fun” – to laugh at people who believe their false creations.

APA uses Tickaroo Live Blog technology. Liveblogs are fact-checked: a tricky process, given the time pressures of liveblogging.

Cat and mouse

Countering misinformation is an arduous task – not because it is hard to find the correct facts and information, Florian says, but because it is difficult to get the get the message across. Fact checks are often rather boring text, very scientific and long.

“So the question is how can we get those facts across. We are completely aware that we cannot reach the people who are very deep in those conspiracy bubbles – there are studies that show these people cannot be reached by fact checks, but you can reach the people standing on the edge who are in danger of falling into the conspiracy bubbles.

“I think we write the fact checks for them; for people who search for information, who search for correctness, [who want to] understand what’s wrong and what’s right about the claim,” Florian says.

Keeping up with evolving threats and artificial technology is “like a cat and mouse game”.

“It’s not a one-man job or even a four-people job. We have four people [on our team] and of course we can’t follow all the developments on our own.”

That is why, Florian says, it is important to collaborate with other fact checkers nationally and internationally.

Fact-checking is a crucial tool to rebuild trust, Florian stresses, predicting that countering misinformation with fact-checking will probably change the role of journalism.

“In the last centuries and years we were known for spreading information to different channels, but I think this role isn’t so important anymore because the internet allows us to have a lot of information coming from different sources,” Florian says.

“So, there’s less necessity to spread information, but there is of course a necessity to verify information.”

  • Watch our wide-ranging interview with Florian below.

Media Makers Meet – Live
We’d like to see you at our upcoming live events in Spain and Portugal
  • Mx3 Barcelona focuses squarely on innovation in media, emphasising creator-led, consumer and B2B media operating in and across media verticals. It takes place from 12-13 March. Click here for more.
  • The FIPP World Media Congress, which we produce on license from FIPP, takes place from 4-6 June in Cascais, Portugal. The magazine media world’s flagship global event will be the 46th edition since it first launched in Paris in 1925. Click here for more