Asa society we have never had so much information at our disposal. With the proliferation of disinformation (information that is false and deliberately harmful to a person, social group, organisation or country), the journalist’s ongoing task to cement themselves as a reputable information source is harder than ever before.

Fact-checking has always been a necessary practice of a good newsroom. The rise of artificial intelligence (AI) has made it harder for journalists to accurately and quickly verify sources of disinformation because ‘fake’ news is becoming increasingly realistic.

There are positive technologies being used by journalists to combat disinformation, such as Google’s Reverse Image Search engine, TinEye, or Yandex. But AI is developing quickly; in May 2019, AI researchers at the Skolkovo Institute of Science and Technology in Moscow revealed their “few-shot” AI system that creates a convincing fake using simply a few still photographs of someone’s face. If newsrooms do not develop strategies to identify malicious uses of such AI, the public risks being fed false information by the very institutions that are meant to fairly and accurately report on current affairs.

Gabon President Ali Bongo’s New Year’s address was labelled as ‘deepfake’ by his opponent Bruno Ben Moubamba after he appeared in public after a period of prolonged illness. This scepticism about the video’s credibility sparked a failed military coup. Although the video is yet to be proven fake, this controversy combines with the viral ‘shallowfake’ of US House Speaker Nancy Pelosi that was retweeted by President Donald Trump, to illustrate the potential for AI technology to be weaponised to serve a political agenda.

The worrying fact is that social media users are not sharing disinformation simply out of ignorance. According to a Loughborough University report which had 2,005 British respondents, of those who had shared news on social media in the past month, 42.8% shared inaccurate or false news. This included 17.3% who shared news they thought was made up when they distributed it.

The report also found that Conservative supporters and those with right-wing ideological beliefs are more likely to share false or inaccurate news and to be reprimanded by others for doing so, whereas Labour supporters, and those who hold left-wing ideological beliefs, are more likely to encounter inaccurate news and to correct other social media users for sharing it.

It is in the interest of democracy that journalists learn to identify and discredit sources of disinformation quickly before they become widely distributed online, as one side of the political spectrum is more likely to be corrected for practicing disinformation than the other. Newsrooms must work closely with social media platforms to refine their content strategies if they are to be successful in recovering at least some of the trust between the public and mainstream media that has been falling for decades.

The democratic threat posed by malicious AI and disinformation calls for a radical overhaul of the newsroom, with a need for a move towards a ‘slow news’ model. This is because the financial structure of the newsroom, based on being first to the scoop, makes journalists especially vulnerable in ‘breaking’ news situations.

The democratic threat posed by malicious AI and disinformation calls for a radical overhaul of the newsroom, with a need for a move towards a ‘slow news’ model.

The reporting of catastrophic or political events could easily become a high stakes game in which malicious agents could commit a new form of terrorism, as we began to see with Russian interference in the 2016 US election. Disinformation attacks would be supported by the fact that the web tends to reward shocking stories with bigger audiences, a phenomenon exemplified by the horrific but very real New Zealand terror attack that spread quickly amongst Facebook users.

Tortoise Media is an early adopter of this proposed ‘slow news’ model, and is committed to offering the news ‘not as it happens, but when it’s ready’. Its name is derived from Aesop’s fable The Tortoise and the Hare, and the proverbial takeaway of ‘slow and steady wins the race’ certainly has an important application to the traditionally fast-paced and frantic newsroom.

Tortoise Media states that the problem isn’t just ‘fake news’ or ‘junk news’, because there is a lot that is good. The problem is that there is such a large volume of information out there; Filippo Menczer’s research has highlighted the challenges of our brains to make decisions about the credibility of information when the streams of content are overwhelming.

If journalists slow down and think carefully about where their information has come from and the range of voices they are listening to, newsrooms will have a reduced but higher-quality output. The more tortoise newsrooms there are, the greater the reduction in noise will be. And maybe then, finally, the public will trust us again.

Would you buy into a slower model of journalism? Or does ‘breaking news’ have to be by its very nature instantaneous and sometimes erroneous?