News

Fact-checkers are fighting multi-headed disinformation

By Nalaka Gunawardene

Michael Ondaatje, the Sri Lankan-born Canadian poet, essayist and novelist, once wrote, “In Sri Lanka, a well-told lie is worth a thousand facts” – and how perceptive he was!

That remark, tucked away in the acknowledgements of his memoirs Running in the Family, first published in 1982, is worth revisiting today as our society struggles with a deluge of fabrications, distortions, conspiracy theories and other abuses of facts.

Is it the rising levels of internet use and especially social media proliferation that has caused this, as some claim? 

For sure, the web allows falsehoods to spread faster and more widely than ever before. But as Ondaatje pointed out, cooking up tall tales and taking liberties with facts seem pervasive in Sri Lanka.

Indeed, many of us seem to straddle two worlds: one of verified facts that are used as evidence for reasoning, and the other based entirely on conjecture, beliefs, or collective imagination. 

For example, sufficiently large numbers of Lankans have uncritically accepted claims ranging from ‘infertility pills’ and ‘resurgent Tigers’ to supposed conspiracies by multinational corporations to ‘poison the Lankan nation’ with highly-toxic agrochemicals. Some of these claims have aggravated communal tensions spilling over into violence (as in March 2018).

So understanding how falsehoods emerge, spread and pollute the public mind is not just an academic pursuit in Sri Lanka. It has implications for the quality of public discourse and strength of our democracy.

For this reason, I have spent time and effort in recent months to try and understand how falsehoods and false narratives affect public policies, voter behaviour, market trends, and even societal tensions in Sri Lanka.

My enquiries were prompted by my background and experience as a science journalist trained to look for data, evidence and informed opinions on complex issues. 

I wanted to go beyond simple, binary (yes/no type) questions and explore nuances of the matter. In that process, I asked questions such as:

Do we as a society assign enough value to facts and, if not, why?

Or is ‘fact-resistance’ embedded in our collective psyche? 

Whose facts are we talking about in a pluralistic society with many contested spaces and contentious issues?

With Lankan politics saturating with spin-doctoring and manipulations of public opinion, can the fact-checking of politicians make a difference to polity or governance?

These are admittedly open and broad questions but necessary. I have not yet found all the answers, but some aspects of the problem have become clearer. Let me share some useful insights.

Insight 1: Information society is a paradox

In the early days of the digital age in the 1990s, when public access internet services began to spread, it was widely believed that it would lead to a freer flow of information and a flourishing of free expression. These did happen, but there were also some unintended consequences.

By the mid 2000s, web technologies had evolved to enable easy self-publication through social media platforms like Facebook, YouTube and Twitter (and comparable Russian and Chinese services). Anyone, anywhere with basic computer literacy and connectivity could now produce content and disseminate it to a potentially global audience. 

This outpouring of self-expression has been a mixed blessing. On the plus side, bloggers, Facebookers, Instagrammers and YouTubers are bearing witness, speaking their minds and having myriads of conversations that were earlier impossible or tightly controlled by legacy media.  

Yet the loss of control by traditional information ‘gatekeepers’ such as editors and broadcast news managers has opened the floodgates to the good, bad and the ugly. This includes peddlers of dubious cures, financial swindlers, hate-mongers, various political manipulators and conspiracy theorists.

Insight 2: Lies travel faster and wider

Again, this is not a new trend. Back in 1710, the Anglo-Irish satirist and essayist Jonathan Swift (1667–1745) noted: “Falsehood flies, and truth comes limping after it, so that when men come to be undeceived, it is too late; the jest is over, and the tale hath had its effect.” 

During the past two decades, however, falsehoods have gained unprecedented speed and reach thanks largely to social media. While fact-checks and other attempts at debunking of falsehoods have access to the same online platforms, they often don’t spread as fast or wide.

The reasons for this are being debated and studied. Is it because falsehoods tap into people’s deep-rooted prejudices, suspicions or fears? Or do some people just prefer comforting lies over inconvenient truths?

One recent, multi-year study by three researchers at the Massachusetts Institute of Technology (MIT) found that false news spreads more rapidly on Twitter than real news does — and by a substantial margin. Their analysis of over 126,000 news stories tweeted by three million people between 2006 and 2017 found false news stories were 70% more likely to be re-tweeted than true stories. It took true stories around six times longer to reach 1,500 people on that platform. 

The study also revealed that fake news was more commonly re-tweeted (i.e. shared) by humans than bots, which are software programs that perform automated, repetitive tasks. [More at: https://news.mit.edu/2018/study-twitter-false-news-travels-faster-true-stories-0308]

The MIT study’s findings may not be directly applicable to other tech platforms or societies, but the generic concern is shared by fact-checkers and others among us who demand hard evidence before accepting any claim. Due to the ease of sharing on both social media platforms and instant messaging services like WhatsApp, too many people uncritically forward unverified information – sometimes without even reading beyond the heading.

Insight 3: Information Disorder requires multi-pronged responses

Societies have historically relied on journalists and subject specialists to discern facts from the rest. Such filtering methods simply cannot tackle the scale, speed and growing sophistication of online disinformation today.

The societal costs are becoming clear. The deluge of lies, propaganda, conspiracies, rumours, hoaxes, hyper-partisan content, and manipulated media is increasingly undermining the integrity of elections, hampering pandemic responses, and threatening rule-based systems of governance. Climate denialism may turn out to be the costliest falsehood of all time.

These trends, in turn, are eroding public trust in the news media, democratic institutions and technical experts.

What is to be done? And how must open societies respond to this challenge without unduly restricting freedom of expression?

Different countries are trying out various strategies. These include: introducing new laws regulations to deal with the really harmful falsehoods; increasingly the supply of authentic information; promoting digital literacy so people become more critical users of information; supporting independent journalism; and fact-checking.

The Information Disorder – the preferred collective term among those studying this phenomenon – has complex socio-political roots and keeps evolving. As such, there is no single or easy solution whether legal or technological. It requires a multi-pronged response. 

Insight 4: Fact-checking is necessary, but not sufficient

One proven strategy in countering falsehoods is fact-checking. Reputed mainstream media institutions are expected to fact check before publication or broadcast, as part of their commitment to professionalism. That rarely happens in Sri Lanka. Instead, we have post hoc fact-checking, i.e. independent researchers or civil society groups verify content after publication in mass media or social media.

In 2017, when I first publicly advocated the need for such a service in Sri Lanka, there was none. Verité Research, an independent think tank, stepped up to the challenge in 2018 by creating Sri Lanka’s first independent fact-checking service: FactCheck.lk.

Since then, several more have emerged. Half a dozen fact-checking services operate in Sri Lanka today – some run by journalists, others by think tanks, or civil society groups. Two are international services covering Lankan content, while the others are home-grown. The government’s Department of Information has also recently launched one. 

Their scope, emphasis, and methodologies vary, as does their outreach. Some probe selected news stories in legacy media. Others monitor social media and investigate dubious or suspicious claims.

Yet one thing is clear: their combined efforts are woefully inadequate. Too many falsehoods keep coming out, and online, nothing ever really goes away (old lies can keep resurfacing for years). Fact-checkers have to be highly selective in what they choose to probe.

Some lies and distortions need to be debunked quickly, before they reach too many people and trigger communal unrest or even violence. During the anti-Muslim violence in Ampara and Kandy Districts in March 2018, and in the days following the Easter Sunday terror attacks in April 2019, various unverified claims of attacks and counter-attacks circulated through social media and messaging apps. Many turned out to be dangerous fabrications. 

On both occasions, the Government ordered temporary blocking of selected platforms and services to contain the problem. Such blocking is a blunt regulatory measure that should be the last resort when all else fails.

Insight 5: Fact-checkers are running an unfair race

Sri Lanka’s growing number of fact-checking entities are run by highly-committed men and women who work hard. Yet they must feel overwhelmed or frustrated by how easily blatant lies or cleverly-disguised falsehoods seem to spread. 

In this never-ending race between facts and its perversions, fact-checkers are like foot soldiers fighting in the trenches, one debunking at a time. Their adversaries get craftier and more sophisticated every passing year: instead of individual falsehoods, the trend now seems to be in creating false narratives.

What makes every fact-checkers job unenviably hard is well explained by the ‘bulls**t asymmetry principle’. Publicly shared in 2013 by Alberto Brandolini, an Italian computer programmer, it says: “The amount of energy needed to refute bulls**t is an order of magnitude bigger than to produce it.”

In recent years, the term ‘bulls**t’ has moved from being a mild expletive to a term used to describe communications without any grounding in facts or truth. It was the American philosopher Harry G. Frankfurt who first wrote a significant academic paper on the subject in 1986. He later expanded it into a popular book titled On Bulls**t (Princeton University Press, 2005) which was on the New York Times best-seller list for weeks.

AI help for human fact-checkers?

Given such formidable odds against which fact-checkers work, are there ways to help them scale up and speed up debunking without compromising quality or accuracy? Is this something where Artificial Intelligence (AI) can be customised to help?

The ICT think tank LIRNEasia has been studying this challenge using Sinhala and Bengali languages. Their research shows that by partly automating the initial detection and filtering work, machine learning systems can indeed enable human fact-checkers to scale up their work.

These findings were shared with fact checker community and others at a dialogue forum in mid-December 2021 that I moderated. The next article will summarise key points emerging from our wide-ranging discussions on fast-tracking fact-checks through human and machine collaboration.

(Science writer Nalaka Gunawardene has been chronicling and critiquing information society for over 25 years. He tweets from @NalakaG)