who gets to judge what we see and say?

On Sunday, Brogan Renshaw’s lifelong habit of reading the news hit a wall. Since Renshaw, 32, was a boy, his dad would unfurl the newspaper and the day would begin.

“I follow The Sydney Morning Herald. I Google it all the time. My general search is just ‘SMH’,” Renshaw says. ”That day the site did not come up. I thought maybe it’s just me.”

It was not.

Google had quietly started to hide local news from Renshaw’s search as part of an “experiment” aimed at showing how much the media depends on traffic from the company. Whatever one thinks of Google, it was a demonstration of the power of the technology giants that run much of the internet. The past fortnight has been full of other examples.

Twitter, YouTube (itself owned by Google), Facebook and its photo app Instagram all banned or suspended US President Donald Trump after the violent storming of the Capitol by a mob of his supporters.

Shopify, which is behind many of the world’s online storefronts, and Stripe, which processes payments, cut off Trump-related accounts. So did Campaign Monitor, a service his campaign used to send email blasts.

In a dramatic move underscoring their control of speech on apps they do not even own, Amazon, Google and Apple took Parler, a free-speech social network favoured by Trump supporters, off their app stores and the internet. “We’re toast,” Parler’s policy boss Amy Peikoff said when the bans were threatened.

Loading

Google’s decision to stifle Australian media on its platform was triggered by a different motivation to the other platforms, which acted to stop the incitement of violence and hate speech. The search giant is at odds with the government and industry over a news bargaining code, currently before the Senate, that would force it to pay for the news content it displays in a manner it thinks unfair.

By contrast, two weeks ago Trump had riled up an insurrectionist mob and directed them to the Capitol building, which they stormed as lawmakers were certifying election results, while Parler hosted graphic threats to murder Black and Jewish people along with police officers.

Many cheered the technology giants’ actions. Others did not. Among them were some who have their political futures yoked to Trump either here or in America.

But there is a third group: people who dislike or even loathe Trump and may even support his ban, but find it worrying the way tech barons atop trillion-dollar companies can act as unappealable global censors. Russian opposition leader and anti-corruption activist Alexei Navalny called it an “unacceptable act of censorship” that would give the green light to regimes globally to do the same and German Chancellor Angela Merkel said it was “problematic”.

These debates have upended traditional political divides.

There are small “l” liberal warriors whose insistence on free speech has crashed into their insistence on free markets, overturning the classic argument that private companies can choose the terms on which they take customers.

NSW Liberal Senator Andrew Bragg is clear. “I’m a markets guy,” he says.

But he is unhappy with the platforms’ power to silence and is pushing for a mandatory code to regulate what content remains online and what is taken down if the companies don’t agree to self-regulate first.

“The reality is they are the railroads of the 21st century. As you know the US imposed the breakup of the railroads,” Bragg pauses. “There are a lot of parallels here.”

Deeming the platforms publishers is not a position the government has moved to adopt. Labor, a party generally much happier with regulation, has taken the opposite approach to Bragg.

“I’m a bit old fashioned, I believe in free enterprise,” said the party’s health spokesman, Chris Bowen. “I believe that Twitter is a free enterprise and should be able to ban whoever is inciting hate speech as they wish.”

More than 50 MPs from all parties have joined a parliamentary group dedicated to making social media safe.

If there is one thing that unites the industry’s critics on both sides, it is a call for consistency and a view that free speech standards chosen by the people through their governments should determine what remains online.

“There needs to be consistency to the way in which these tech companies moderate content, with transparent rules and guidelines for everyone,” Technology Minister Karen Andrews said this week.

Loading

“It concerns me that an outcry by prominent people or certain groups on social media may dictate their decision-making when there is incredibly vile, hateful and dangerous content that frequently goes unchecked.”

Examples are plentiful. Some are relatively constrained. A conspiracy theorist who falsely claimed Australian politicians were paedophiles remained on Facebook until Victorian Nationals MP Anne Webster successfully sued. The poster has now shifted to YouTube. Some are not. In perhaps the worst case, Facebook failed to stop its platform being used by the Myanmar military to incite genocide against the Rohingya in the late 2010s.

It has since admitted to initial failures, banned some top military figures in Myanmar and said it was proactively taking down misinformation in the country.

Others in the sector admit the status quo has not worked well, even if Trump’s ban was the right result. Twitter co-founder and chief executive Jack Dorsey defended kicking the President off the service that has defined much of his presidency on Thursday, Australian time, because of an imminent risk of violence but said it marked larger problems.

“I feel a ban is a failure of ours ultimately to promote healthy conversation,” Dorsey said.

“Having to take these actions fragment the public conversation,” he said. “They divide us. They limit the potential for clarification, redemption, and learning.”

The idea that if people did not like Twitter they could simply go elsewhere had also been thrown into question, Dorsey said: “Parler was gone, although it could return, and other platforms had also banned Trump.”

But the agreement that the technology giants’ power, inconsistency and opacity is a problem is not a consensus on a solution.

One model proposed by Dr Webster is to deem the platforms local publishers, making them easier to sue for the content they host just as television stations can be responsible for the comments of their talking heads.

That would encourage Facebook and others to take down content that already breaches Australia’s discrimination, defamation, and classification laws faster because their money, rather than just their users, would be on the line.

In Dr Webster’s case, the MP won $875,000 for defamation but the conspiracy theorist has not paid up and is living in New Zealand. Dr Webster cannot pursue Facebook for that money.

“My husband and I are still paying for court costs, $160,000, and it’s still not finished,” Dr Webster says. “We’re still paying to have justice, who can afford that?”

One response from the social media and search giants is that they are neutral platforms, not publishers, because they merely serve up what their users provide.

That argument became harder to maintain this week when Google on Wednesday confessed to running its “experiments” that hit users like Renshaw, temporarily shunting local news publishers from their normal position in search results.

It is only because Renshaw runs his own digital agency, Firewire, that he thought to capture the Herald’s absence before it reappeared.

Along with Nine-owned publications including the Herald and The Age, News Corp papers such as The Australian were targeted.

A Google spokesman says the changes, which will run until next month, were intended to measure how much value Google provides news organisations and vice versa as the company lobbies against the government’s proposed bargaining code.

Google is not against the concept of a code to support the media but wants a different arbitration model and says the current design would “break” its search engine.

It has not said how many Australians’ news access is affected, but disclosed there are a “few” experiments running, each affecting about 1 per cent of search users, which given the company has about 95 per cent of the market, could total hundreds of thousands of people.

“It’s a threat by Google frankly,” says Jake Goldenfein, a law and technology researcher at the University of Melbourne.

It also shows, Dr Goldenfein says, that far from being a neutral platform without oversight of what it spreads, Google can rapidly change what searchers see based on its political and commercial aims.

Loading

The mainstream social platforms are not free-for-alls like Parler. Facebook, Google and Twitter do have moderation teams and policies to police inappropriate content according to their terms of service, but critics say they are spread too thin.

“It happens at extreme scale with underpaid and undertrained workers exposed to traumatic material,” Dr Goldenfein says.

Technology site The Verge revealed in 2019 how contract moderators for Facebook in America started to experience panic attacks after watching one too many videos of a slaying while others came to believe the conspiracy theories they worked to remove, proclaiming the world to be flat.

(Facebook has since agreed to a settlement for the moderators’ post-traumatic stress disorder, stopped using the provider investigated by The Verge and committed to making the work safer for its employees.)

Queensland University of Technology academic Terry Flew, who led a major review of Australia’s traditional media classification laws a decade ago, says the idea of the government stepping into moderate banned content itself is a bad idea.

For one, it has not gone well historically. Assessing George Orwell’s 1936 novel Keep the Aspidistra Flying, one Australian government censor wrote that it had “No literary merit, and I consider it indecent. I would ban”.

Customs officers were duly instructed to seize any copies at the border.

For another, the speed, scale and linguistic challenges of moderating content across countries and languages mean the work should stay with the platforms, Professor Flew says, but not without oversight from the government on the standards moderators apply.

Loading

“There should be a role for government and in an interesting way the tech companies would probably welcome that,” Professor Flew says. “Because at the moment they face the problem that they are unacknowledged legislators.”

That dovetails with a broader question. What to do about the content that might be false, but does not fall foul of Australia’s current laws, such as Sydney Liberal backbencher Craig Kelly’s claim that wearing masks is “child abuse” or debunked posts from Queensland Nationals MP George Christensen’s about purported US election fraud?

“To have members of Parliament spreading false and misleading information is a shocking state of affairs,” former Liberal prime minister Malcolm Turnbull told Nine News. Labor wants Prime Minister Scott Morrison to reprimand the backbenchers for spreading their quackery but he has so far not done so, instead saying Kelly and Christensen are free to speak their minds.

Both men are backbenchers with no real say on government policy, but the situation in the United States shows how disinformation about things like an election can spiral into widespread belief that false claims are true.

“A platform like Facebook is showing you material for two reasons. Purpose one is to show you ads and purpose two is to show content to keep you on the platform to show you ads,” Dr Goldenfein says.

That allows engaging but inflammatory content to rise to the top.

Loading

Chris Cooper, from Reset Australia, an advocacy group that tries to stop technology damaging democracy, says the first step on this front is to understand what is really going on inside the platforms, where the companies have a bird’s-eye view while others only see a black box.

He wants the platforms to be forced to tell governments and regulators what their top content is on key topics, like the coronavirus, in ”real-time” so that civil society can identify if there are problems with what their algorithms are promoting.

“If we’re going to roll out a vaccine they should have more info and insights in real-time about what Australians are consuming,” Cooper says. “It becomes less about freedom of speech, though that should be protected, and more about what is amplified and what isn’t.”

It is hard to know what principles should guide that effort but Google’s move to bury reliable mainstream news is not an auspicious start.