Fact checking has been described as cleaning up a sewage spill with a teaspoon. What happens to truth when someone takes the teaspoon away?
When Donald Trump fronted his first debate of the US presidential campaign against then-Vice President Kamala Harris, back in September last year, he made what has become a notorious accusation: migrants in a small Ohio town were eating the local pets.
"A lot of towns don't want to talk about it because they're so embarrassed by it," then presidential candidate Trump said. "In Springfield, they're eating the dogs. The people that came in, they're eating the cats. They're eating — they're eating the pets of the people that live there."
This assertion by now-President Trump, that Haitian immigrants were abducting and eating their neighbour's dogs and cats, was rapidly fact-checked by multiple news outlets and eventually awarded "Lie of the Year" for 2024 by PolitiFact, an organisation that ushered in a new era of fact-check journalism when it launched in 2007.
From Time magazine's pioneering research department in the 1920s to the New Yorker's renowned and rigorous fact-checking department, verifying facts is as old as journalism itself. The rise of the 24-hour news cycle, ubiquity of social media and subsequent spread of online misinformation and disinformation have since demanded more from journalists and editors. It has forced newsrooms to improve their internal verification processes but has also led to the proliferation of specialised fact-check outfits like AAP FactCheck in Australia or AFP Fact Check, which now has a network of 150 journalists across 26 languages publishing checks on everything from Russia's invasion of Ukraine to COVID-19.
Among the thousands of fact-checks published in the past year fact-checkers in Australia and abroad have debunked a wealth of manipulated footage in the Israel-Gaza conflict, found Labor misrepresented the Coalition's stance on free GP visits, concluded Opposition Leader Peter Dutton was wrong on tourist visas, discredited the claim Kamala Harris wanted to take away Americans ability to "eat red meat" and investigated whether planes are actually crashing more often (they're not).
The era of 'fake news'
The political ascendancy of Trump heralded a golden era for fact-checkers. Throughout, Mark Zuckerberg's social media giant Meta played a significant role in the fact-checking industry.
In 2016, soon after Trump's first election to the US presidency, Meta launched its program paying independent fact-checkers to flag and debunk disinformation on its products, mobilising a global battalion of people fighting falsehood's on the world's most popular platforms.
It wasn't just Meta's partners who expanded their fact-check operations during this period but legacy media outlets. During the first Trump presidency The Washington Post's fact-checkers declared he had made 30,573 false or misleading claims. By 2017 "fake news" was named the word of the year. In August last year the then-presidential candidate almost backed out of an interview once it became clear he would be fact-checked live and then in October cancelled an interview with American 60 Minutes after his campaign complained the program would fact-check the interview.
Between 2014 and 2021 the number of international third-party fact-checkers grew tenfold but after eight years of continued growth, research from the Reporters' Lab, a centre for journalism research at Duke University, found that the number of active fact-checking projects across 111 countries dropped for the first time in 2023 and then again in 2024.
Then, in the wake of Trump's recent election win, fact-checkers were dealt another blow. Meta, which owns Facebook, Instagram and Threads, promptly scrapped the teams of professional and third party fact-checkers in the United States.
"It's time to get back to our roots around free expression," Meta founder and CEO Zuckerberg declared in January.
And with that fact-checkers were replaced with a crowdsourced "community notes" system similar to that on Elon Musk's X. Instead of using accredited fact-checkers to debunk and flag misleading or inaccurate content that should be marked with a warning label, this model relies on other social media users to add context or caveats to a post.
Right now, the roll-back applies only to the US but Meta funds more than 100 fact-checking organisations across the world and has not ruled out further partnership terminations.
The social media giant's most recent report, via Australia's code of practice on misinformation and disinformation, found that in the 2023 calendar year it took action on more than 9,700 pieces of content for violating misinformation policies. Warnings were displayed on more than 9.2 million distinct pieces of content on Facebook based on articles raising concern that were written by its third-party fact-checking partners in Australia at the time: AAP FactCheck, Agence France-Presse and RMIT FactLab.
Zuckerberg has promoted the about face as a win for "free expression".
"Governments and legacy media have pushed to censor more and more but now we have the opportunity to restore free expression, and I'm excited to take it," Zuckerberg, who once proudly told US congress Meta's fact-checking program was "industry leading", said.
Meta's decision was criticised as 'spineless and opportunistic', condemned as a move that will disproportionately harm marginalised communities and described as a change that would encourage the "exact same surge of hate, disinformation and conspiracy theories that caused January 6". Critics of the move have said it was made in an effort to curry favour with the incoming US president, who has been routinely fact-checked by multiple third-party outfits and publications.
Asked directly if Zuckerberg's decision was a response to threats Trump had made to him in the past (he once threatened to jail the Meta CEO), the president responded: "Probably. Yeah. Probably."
The growing politicisation of fact-checkers
In November Brendan Carr, picked by Trump to head the Federal Communications Commission, sent an email to the CEOs of Meta, Apple, Alphabet (Google's holding company) and Microsoft. He accused them of participating in a "censorship cartel" and singled out fact-checking.
This suspicion of fact-checking is not only shared by members of Trump's administration but by his voter base. A 2019 study found seven in 10 Republicans say fact-checkers tend to favour one side, compared with roughly three in 10 Democrats. In the lead up to the US election those on the left of politics criticised CNN for declining to fact-check Trump in real time during his debate with Joe Biden while those on the right criticised (the American) ABC News moderators for fact-checking that network's debate.
In the wake of Meta's decision, 130 fact-check organisations wrote an open letter to Zuckerberg reiterating the "strict non-partisanship standards" Meta's partners had to attain through verification by the International Fact-Checking Network, which requires annual verification, including independent assessment and peer review. They disputed the tech billionaire's accusations of censorship as Meta never granted fact-checkers the ability or authority to take down content or remove accounts.
The head of the network, Angie Drobnic Holan, said at the time that Meta's move was made "in the wake of extreme political pressure from a new administration and its supporters".
Holan is a former editor-in-chief of PolitiFact — one of the first dedicated fact-checking groups — and was part of the Pulitzer Prize-winning team of journalists who examined claims made by politicians during the 2008 US presidential campaign. She says there has been growing political pressure on fact-checkers, particularly over the past 15 years.
"The most disturbing trend is that activists, political parties and elected officials have tried to discredit fact-checking as an activity in and of itself because it doesn't align with their preferred narratives," she tells the ABC.
"This includes direct attacks alleging bias or political motives, attempts to discredit fact-checkers as incompetent or as censors and overarching efforts to undermine public trust in independent journalism and verification."
What's happening in Australia?
Fact-check journalists and misinformation experts in Australia have testified that their work feels increasingly politicised while one of three third-party fact-checking outfits here has shut up shop. RMIT FactLab recently confirmed it has ended its third-party fact checking work. In June last year the ABC ended its separate and seven-year collaboration with the university called the RMIT ABC Fact Check unit. AAP FactCheck recently confirmed it will continue to provide checks for Meta for at least another 12 months while international news wire AFP told the ABC the outlet was also committed to the program for this year in Australia.
"Fact-checkers in Australia and elsewhere have unwillingly become part of the culture wars and a target for people who want to make the claim that free speech is being quashed," William Summers, former chief fact-check journalist for AAP FactCheck, says.
During the Voice to Parliament referendum campaign RMIT FactLab drew criticism from Sky News Australia, Liberal senator James Paterson and conservative think tank the Institute of Public Affairs during the Voice to Parliament.
Andrea Carson, La Trobe University's professor of political communication, says in 2022 her team did survey experiments on trust in fact-checkers in Australia and found there were "high levels of trust in Australia for fact-checking" and, at that time, "very little difference between the brands".
"Fact-checking tends to be more highly regarded by those on the left side of politics than the right in the US and we are seeing some of that politicisation and polarisation of fact-checkers starting to happen in Australia but it became most prominent during the Voice to Parliament referendum," Carson tells the ABC.
"We did another study during the Voice to Parliament referendum, also on fact-checking and asking about trust. We saw a noticeable dive for RMIT FactLab and that came off the back of that concerted campaign against RMIT FactLab. We saw trust had fallen for that brand for those on the right side of politics and particularly those who were consumers of Sky News."
Carson stressed that while Australia was experiencing "elements of politicisation" when it comes to fact-checking there was still a high level of trust, and it was "nowhere near the extreme of what is going on in the US".
"Also, when trust does fall like it did during that concerted campaign for that particular brand, we don't know how long that lasts unless we go back and study it," she says.
The Conversation conducted research with global media monitoring company Meltwater and found of the most popular posts in Australia about the Voice in a three-month period leading up to the referendum, an article by Sky News Australia accusing RMIT FactLab of working with Meta to "censor" the Voice debate came in second.
Concerns about fact-checkers have been raised in Australia's parliament by politicians over the past few years.
Gerard Rennick, in 2023 questioned bureaucrats bureaucrats from the Australian Communications and Media Authority about "who gets to fact check the fact-checkers" and grilled the ABC chair on a fact-check involving a vaccine death. Malcolm Roberts said all three of Australia’s third-party fact-checking organisations exercised "consistent bias" and late last year Matt Canavan accused AAP FactCheck of “effectively silencing political arguments” during the Voice debate.
Meanwhile fact-checkers and misinformation academics have raised their own concerns about the increasing politicisation of their work.
"During the Voice referendum there was so much misinformation and disinformation being pushed around but because it was a politicised debate, fact-checkers were being accused of bias," Summers remembers.
He says not only did no one in the team embark on a fact-check to make a political point, "the opposite happened to a certain extent". He says fact-checkers struggled to find more false claims from the Yes campaign "because there were so many false claims from the no side it dominated the debate".
These false claims included that ticks in the No box would be counted as a Yes; Pauline Hansons's assertion that politicians would have to consult the Voice about private members' bills; former boxing champ Anthony Mundine’s claim that Indigenous people would become British subjects if the Yes vote won at the referendum and and Dutton’s claim that the prime minister had never mentioned the Voice before the election.
A cyber security report by global intelligence research organisation Recorded Future revealed concerted efforts by far-right groups and an army of bots to spread false information undermining the Voice to Parliament. A few months before the vote Meta shut down 9,000 Facebook and Instagram accounts run by a group linked to Chinese authorities that, among other things, had been spreading misinformation on the Voice.
Summers believes most people are unaware that some foreign countries are "pushing to pollute Australia's information space".
He described a fact-check he did last year for AAP in which he looked into posters, claimed to have been created by Amnesty International, protesting Israel's participation in the 2024 Olympics and found in fact Iran's state media had published the computer-generated images of Israeli athletes brutalising Palestinians.
"A lot of this misinformation and disinformation is not just naively shared, some of it is very intentionally shared by people who don't have Australia's interests at heart," he says.
The removal of fact-checks comes as Meta revamps a program that rewards creators with bonuses based on views and engagement. A former Meta data scientist has warned this could create a welcome environment for those spreading "viral hoaxes for profit".
There were usually two reasons people spread falsehoods, Summers says: for ideological gain or for financial gain.
"Most social media platforms now pay out for viral content either directly when you get over a certain number of views or indirectly because you are building your own platform and one easy way to do that is to produce content which is surprising but a lot of surprising content either misses context or is just completely untrue," he says.
"We are more likely to share the surprising unverified content over verified facts and uncertainty."
Is there a role for fact-checking in 2025?
Carson argues that fact-checking was good at identifying misinformation in the first instance but is not a "standalone solution" and has its limits.
"It is a fairly consistent finding that you can't really use fact-checking to move viewpoints of a political nature, and we found that in our own research," she says.
"When people have a set idea about something and it is fact-checked, it doesn't move their opinion of it because they have motivated reasoning, which is usually politically motivated, for staying with that particular idea."
One of Carson's studies found that even though people trusted a fact-check and "accepted it had done its job", it didn't always stop them spreading the political misinformation.
"A third of our respondents just went ahead and spread it anyway both on the left and the right of politics and they gave different rationales for that," she says.
The study used a video of Scott Morrison that had been falsely labelled to say the former prime minister had blamed flood victims for "weaponising their trauma" against the government.
"What we found is that people on the left would share that, even though they knew it had been fact-checked and they did so because some of them didn't like Scott Morrison," Carson says.
She says they justified their actions because they believed it was the kind of thing that Morrison would say even if he didn't say it in this instance.
"Those on the right shared it to highlight the kind of things they thought people on the left would spread and to call it out," she says.
Holan is clear on what the process does — "provides people with reliable information" — and doesn't do.
"It doesn't promote political agendas, win or lose elections, or convince people to change their political values or beliefs," she says.
"When people disagree, it's not necessarily because they have the wrong facts. Sometimes they need to persuade each other and find compromises. That's a political process that goes beyond fact-checking, but facts can help people make a good start on more honest dialogues."
Holan says the priority for fact-checkers in 2025 was to do work that "preserves reality and resists false narratives".
"Fact-checkers need to do the day-to-day work that helps the public understand that facts and evidence do matter, and that reality will eventually assert itself over lies," she says.
The impact of community notes has been mixed.
As ABC News has reported, a study by University of California researchers published by the American Medical Association found community notes provided entirely accurate information about COVID-19 vaccines 97.5 per cent of the time. Another study found community notes increased the probability that the author deletes the original post by 80 per cent.
Ironically, an analysis of almost 1.2 million notes written in 2024 and released earlier this year found users frequently relied on fact-checking organisations when proposing community notes.
However, a study from academics in Luxembourg, Melbourne and Germany published late last year found no evidence that the introduction of community notes on X/Twitter significantly reduced engagement with misleading posts on the platform. The authors concluded the system might be "too slow to effectively reduce engagement with misinformation in the early (and most viral) stage of diffusion".
Carson says there needs to be more peer-reviewed research on how effective the community notes model is.
"The benefits of third-party fact-checking was that it was professionals that usually have training in journalism and it had a degree of accountability because the fact-checkers, at least those used by Meta, were registered with the International Fact-Checking Network and had a set of principles that emphasises political neutrality," she says.
What does the future of fact-checking look like?
Summers describes fact-checking as trying to clean up a sewage spill using a teaspoon.
"And for every teaspoon of information you pull out and fact-check someone is pouring a full tanker back into the ocean," he says.
"Fact-checkers in Australia and elsewhere have all been making significant moves to address some of the more structural issues around educating where misinformation comes from, who is responsible for it and increasing awareness of AI-generated content, and trying to improve awareness of what a good quality study is compared to a poor study written by someone with skin in the game."
Summers says the Meta decision marks a victory in the war against fact-checking but the legacy of fact-checkers will live on in newsrooms as it is now "ingrained" in so much of what journalists do every day.
"Instead of fact-checking teams I think the concept of writing articles with the purpose of checking facts is much more central to what news organisations do now," he says.
Summers is also glad to see publications that both deliver stories exposing falsehoods and also "tell people how to spot misinformation or an AI-generated image", noting the work of teams like BBC Verify and ABC News Verify.
Carson would also like to see even more in-house fact-checking by journalists and more "active adjudication".
"When a false claim is made by a politician instead of 'he said, she said' coverage and giving equal weight to a known false claim, active adjudication is when the journalist calls it out there and then and gives the reason for it rather than giving false balance," she says.
Fact-checking offers an important opportunity to communicate verified and credible information which is important for decision making. This is particularly so when it comes to political choices, Carson says.
"Political choices underpin a healthy democracy and help people make considered informed choices, then they decide who is going to represent them at the ballot box," she says.
Holan says Meta's actions are in response to the politics of the day in the United States "rather than an authentic indictment of fact-checkers' work". She believes the decision is a "temporary victory" for those who want to see fact-checking become extinct.
"Surveys show that the public wants to have online information moderated and that they don't have a lot of patience for navigating through torrents of false news," she says.
She says social media companies that aren't run simply as political projects, like X, will continue to grapple with the need to provide positive user experiences. That includes cultivating reliable information.
"Fact-checkers were here before the Meta program, and they'll be here even if the Meta program goes away."
Credits
Words: Gina Rushton
Production: Catherine Taylor
Images: Lindsay Dunbar
