In the Silicon Valley enclave of Menlo Park, about 50 kilometres south of downtown San Francisco, a vast cluster of unremarkable buildings, fronted by extensive car parks, overlook salt ponds on the barren reaches of the South Bay.
Until you reach a giant thumbs-up “like” symbol at the front of one entrance – and the tourists taking selfies in front of it – there is not a whole lot to suggest that this is the headquarters of Facebook.
The social media giant’s campus resembles some type of anonymous government facility; fitting for an entity arguably more powerful than many countries, and currently embroiled in an international espionage scandal.
In the newest of these buildings, designed by famed architect Frank Gehry, there’s a clearer sense that this is where one of history’s most successful creative companies resides.
Striking artwork adorns the walls, and elaborate staircases connect the floors, one of which is the world’s biggest open-plan office.
The nomenclature for meeting rooms is suitably nonconformist – one I was ushered into was named after a quote from the Hollywood film The Princess Bride (Hello. My Name Is Inigo Montoya. You Killed My Father. Prepare to Die).
At Menlo Park, people are working on all manner of weird and wonderful things: Aquila, Facebook’s solar-powered drone designed to deliver internet to remote regions; Oculus Rift, its virtual-reality headset business.
But the team getting the most attention inside the company at the moment, and from the media and politicians, is involved in a much more daunting (and gruelling) task: eradicating fake news from Facebook.
“It’s not the most glamorous job,” says Sarah Su, a product manager on Facebook’s News Feed team.
It’s November, and Fairfax Media has been granted access to Facebook’s inner sanctum, at a time when the social media monolith is under fire for failing to stop the spread of fake news.
“We’re under a lot of scrutiny,” she says. “We get yelled at a lot. And we don’t really get credit for the progress we have been making. But it’s an important problem, and it’s what keeps me up at night.”Changing the world
Tech companies often like to talk about changing the world.
In its few short years of existence, Facebook has certainly achieved that. Just not necessarily in the way its creators would have envisaged.
The social network is a cultural phenomenon that has revolutionised the way humans communicate, and a business behemoth that has totally up-ended the global advertising industry.
In a little more than a decade, the service founded by Harvard drop-out Mark Zuckerberg in a college dorm room has become a $US500 billion corporate powerhouse, used by billions of humans each day.
Yet, since last year’s US presidential elections, when misinformation spread like wildfire on the platform, the narrative surrounding Facebook has taken a dramatic turn.
Politicians from both the left and the right, and from many corners of the globe, are now gunning for the company.
So too are regulators, including ‘s competition watchdog, the ACCC, which announced an inquiry into Facebook (and Google’s) advertising market power in early December.
Even some of Facebook’s earliest backers and staunchest supporters have turned on it.
Chamath Palihapitiya, Facebook’s former president of user growth, now a billionaire venture capitalist and part-owner of the Golden State Warriors NBA team, this month said he had “tremendous guilt” over the rise of the platform, which he said was “ripping apart the social fabric of how society works”.
Before him, the start-up investor Sean Parker, the company’s first president (who was depicted by Justin Timberlake in the film Social Network) recently raised concerns about how Facebook exploits human psychology, and the impact it is having on today’s children.
After initially dismissing the idea as “crazy”, Zuckerberg finally acknowledged in September that fake news on Facebook might have affected last year’s US election outcome, which, to the surprise of most pundits, saw Donald Trump secure the keys to the White House.
Facebook also recently acknowledged that it unwittingly sold ads to Russian operatives during the election, and that those ads might have reached as many as 150 million people.
Zuckerberg is now outwardly projecting a determination to solve the problem.
The question is whether he’s left it too late.No easy fix
To understand the magnitude of the challenge Facebook is confronting, you need to understand how the social media platform works.
When you log into Facebook, or open it on your phone, a secretive algorithm decides which baby photos, wedding pictures or news stories you actually see, filtering out everything else.
“The Facebook news feed is inherently personalised, to show what’s most important and meaningful to you” Su explains.
The goal is simple: to get you to spend more time on Facebook, so Facebook can serve you more ads.
No-one really knows how the algorithm works, besides that it surfaces things you are likely to engage with (read, click, like, share).
That could mean anything from an image of a dress of debatable colour, to a video of people pouring buckets of cold water on their heads, to news stories confirming your biases, to posts designed to look like news stories, but with little basis in fact, that confirm your biases.
The way Facebook is structured has, at the very least, served to reinforce “filter bubbles” that are polarising Western societies.
At worst, it has helped weaponise and amplify misinformation on an unseen scale.Going viral
Virality on the internet existed long before Facebook (remember email forwards?).
But the rise of the gigantic social media platform – which boasts that it attracts 1.4 billion users each day – has made it possible for more things to go even more viral than ever before.
This dynamic has been deftly exploited by emerging media brands (think BuzzFeed or, in a local context, The Betoota Advocate) and start-ups (think online mattress sales companies), to get content in front of as many people as possible, for as little cost.
But it also has been exploited by people with more nefarious aims.
Shadowy content farms inside the US, and in locations far away such as Macedonia, figured out they could seed inflammatory posts with little basis in fact into Facebook. They would be picked up and shared within certain communities predisposed to believing them, be amplified by the algorithm, and ultimately appear in the news feeds of millions.
For a small content farm, employing one or two people, churning out high-traffic posts and monetising them through cheap ads outsourced to third-party exchanges, this can be a lucrative little endeavour.
That’s why Facebook believes that much of the deliberately false news that has spread on its platforms (which it says is relatively small compared to the real news and information on there) is actually financially, rather than ideologically, motivated.
Facebook is working to remove the financial incentives that motivate content farms to propagate fake news.
“We are really trying to understand what is the difference between good engagement, that actually indicates that you, the reader, are getting value,” says Greg Marra, director of product management on the News Feed team, “and spammy engagement where someone is trying to get the most distribution … because they kind of want to trick you”.
Facebook has tweaked its algorithm to punish clickbait (posts with misleading headlines), and to prevent pages and posts that link out to low-quality web pages featuring intrusive (and sometimes even malicious) ads from generating traffic.
But it freely admits that solving this problem is no simple feat.
“We know that we are working against adversaries that are really creative and highly motivated,” says Su. “So we have to be as creative”.The fundamental problem
Regardless, Facebook’s fake news problem might be a much more fundamental one.
The company’s own a submission to a parliamentary inquiry earlier this year hinted at why.
“The internet … gives everyone the potential to act like a publisher and/or journalist,” the submission reads.
“With just a few clicks, anyone with a smartphone can capture and share breaking news or participate in newsworthy conversations.”
When you create a platform where anyone can publish anything, and absolve yourself from any responsibility from what gets published, that platform is going to be abused.
Of course, people in free societies had been able to say what they wanted long before Facebook arrived.
But they now have access to a mass distribution mechanism to get their views in front of many more people.
A recent experiment Fairfax Media reported underscores this reality.
In the space of a few minutes, and at a cost of just $85, a Sydney man was able to get a misleading Facebook post in front of 10,000 people in Ohio, a crucial swing state in the US.
This illustrates a critical point: much of the most damaging misinformation on Facebook happens through the platform’s front door.
Russian trolls were able to target millions of Americans with ads on divisive issues such as illegal immigration, gun rights and race, simply by setting up pages on Facebook and using the company’s marketing tools to target users that might be sensitive to those issues.
In one particularly disturbing example, Russian trolls set up duelling rallies (one pro-Muslim, one anti-Islam) on the same day, at the same location, in Houston, Texas.What Facebook is actually doing to fix it
Zuckerberg told analysts on the company’s most recent earnings call that Facebook would invest more heavily to thwart fake news, hiring 10,000 people (mostly external contractors) to work on security and safety.
These steps would materially affect Facebook’s profits, he said.
Investors don’t seem too fussed.
“Rather than speaking to the financial community [the announcement] was probably more for the media and Washington DC,” Macquarie Capital tech analyst Ben Schaecter said in a note.
Facebook shares are still within striking distance of record highs, having more than quadrupled since listing in 2012, and business is still booming.
“It’s likely not a coincidence this announcement occurred the same day as members from the top social media companies are testifying before Congress, with potential regulation on the table,” said Schaecter.
The News Feed team is still testing various improvements to ameliorate the problem.
These include new alerts designed to warn people when stories they are about to share are questionable; providing users with easier access to different perspectives and information on the topics they are engaging with.
The company also says it is augmenting the work being done by third-party (human) fact checkers, with machine learning. Once a fact-checker marks a story as false, data from that story is used to inform Facebook’s models, so it can detect future false stories more accurately and quickly.
Once something is determined to be false, Facebook says it reduces future impressions for that post by 80 per cent.
Whether Facebook’s efforts will be enough to resolve the problem remains to be seen.
China reportedly employs millions of people to censor the internet in that country for the 730 million of its citizens who use it.
While no-one is suggesting Facebook should go to those lengths, it makes the 20,000 people it will have devoted to fact-checking and security, for its 2 billion strong platform, looks skimpy by comparison.The only way out of this mess
On Wall Street there are also doubts about whether its efforts will actually succeed.
“I’m sceptical that they will do enough,” Brian Wieser, a New York-based analyst at Pivotal Research, tells Fairfax in an email, “not least as their adversaries will continue working at finding ways to overcome Facebook’s defences.”
Investors won’t be spooked by the problem until it starts to hurt Facebook’s business.
At the moment, there is little evidence of that happening. To advertisers, the ability to serve an ad to almost anyone on the internet, the ability to target who sees the ad based on thousands of parameters, all for a relatively small sum, is proving irresistible.
However, this could change, and quickly.
The ACCC investigation, which has information-gathering powers, will be watched around the world, particularly in light of recent controversies surrounding measurement of the effectiveness of Facebook’s ads.
In the US, recommendations or even legislation from congressional committees investigating social media’s role in US elections, and possible ties between the Trump administration and Russia, could also pose a risk to the company.
At its Menlo Park campus, Facebook will continue to try to make inroads on the problem. The question is whether that effort will ultimately prove futile.
“I don’t know that we will ever be able to drive it [fake news] to zero,” says Su. “But we are going to keep trying”.
John McDuling’s trip to California was funded by an ACS/National Press Club journalism award.