Facebook Admits It Can't Control Facebook Live -- Is This The End For Live Streaming?
In the aftermath of Christchurch has come the admission from Facebook that they can’t control Facebook Live. There’s too much content in general but not enough abhorrent content in particular to properly train their AI. And relying on users to report real time infringements has proven wholly inadequate. Which means that the hypothesis for the company, for politicians and for regulators is very simple. Is it damaging to the public interest to provide a broadcast platform for extremists, for murderers, for the vulnerable, for the suicidal? And, assuming so, if you can’t categorically prevent such incidents (of any scale) from being broadcast live in a way that would be prohibited on mainstream media then what’s the public interest in leaving the system as is?
There was always a likelihood that enabling the population of the world to broadcast live, whatever and whenever they wanted, to Facebook’s global audience might need a rethink. And whilst Christchurch has amplified concerns many times over, there have been other reports of live murders, suicides and violence on the platform. The issue for Facebook is that this now coincides with escalating pressure on their core business from the Cambridge Analytica scandal and the alleged corporate cover-up. Something will have to give to appease politicians and an increasingly skeptical public, and, given the choice, Facebook Live in its current guise might well be it.
When Facebook CEO Mark Zuckerberg published a blog post entitled 'a Privacy-Focused Vision for Social Networking' earlier this month, its intent was to ease concerns about the company's focus on integrated messaging. "I believe the future of communication will increasingly shift to private, encrypted services," he said, "where people can be confident what they say to each other stays secure and their messages and content won't stick around forever."
Three weeks later, Zuckerberg's "simpler platform that's focused on privacy first" seems entirely at odds with an unregulated live-broadcasting service that cannot cope with its (moral if not legal) obligation to censor what is being transmitted and viewed. The pivotal moment came at 1.40pm on Friday 16 March, at the Christchurch city mosque. According to Facebook, the events streamed over Facebook Live were viewed less than 200 times in real-time and then by a further 4,000 people afterward, before the company pulled some 1.5 million uploads from the site.
Facebook's challenge is that this has taken the debate around social media content regulation to a vastly different level. The company has all but admitted that every fail-safe system failed. the content was neither spotted nor reported. Worse, the argument can be made that the availability of a live broadcast medium will become a core focus of future attacks. And that cannot be allowed to happen. If the fail-safes don't work, something else needs to be done instead.
The end of the age of sharing?
"People should be comfortable being themselves," Zuckerberg had written thirteen days before Christchurch, "and should not have to worry about what they share coming back to hurt them later."
Ironically, five years ago Zuckerberg had said that "in five years most of [Facebook] will be video". And Facebook Live's webpage still entreats users to "broadcast to the largest audience in the world with the camera in your pocket." There's a contradiction here with the shift in strategy, and so decisions need to be made. Facebook may well have to start debating the 'least worst' option for what it can sacrifice as calls for the material regulation or even the break-up of social media gather a tailwind.
Until now, the company has fended off such pressure. But in recent days their integrity has been questioned yet again following revelations that staff knew about Cambridge Analytica well ahead of when they 'officially' found out. With more than an eye on business KPIs and its core data trading revenue streams, Facebook Live with its 'uncensored' broadcasts, its inherent risk of reputational damage, and its playing into the hands of regulators may well have to fundamentally change or even go altogether. Facebook was approached for any comments on calls for live streaming to now cease - there was no response at the time of writing.
"We understand there are a lot of tradeoffs to get right," Zuckerberg said in his blog, "and we're committed to consulting with experts and discussing the best way forward." Fortunately, the company has now been presented with an opportunity to do the right thing. So what happens next?
The essential contradiction
Facebook Live soft-launched in 2015 and was fully up and running in 2016, well before the data scandals of 2018 came to light. Since then, there have been incidents of violence, suicides and killings. "We recognize that the immediacy of Facebook Live brings unique challenges," a company blog post acknowledged, shortly after Christchurch. "We use artificial intelligence to detect and prioritize videos that are likely to contain suicidal or harmful acts, we improved the context we provide reviewers so that they can make the most informed decisions and we built systems to help us quickly contact first responders to get help on the ground."
But it isn't working. There is a growing list of inappropriate broadcasts. And the nature of the social media obsessed world we now live in means this is a one-way street. There has always been an uneasy relationship between the media and extremists. Reporting the news without playing into the hands of the perpetrators. The U.K.'s head of counter-terrorism policing called this out again this week, attacking the mainstream newspapers for prompting a far-right agenda under the umbrella of free speech. But, even so, there is still the ability to hit the off switch. But not if the extremists are themselves "broadcasting to the largest audience in the world". And putting in place systems to counter this at Facebook scale is currently an unsolvable problem.
"Many people have asked why artificial intelligence didn’t detect the video from last week’s attack automatically," the Facebook blog post continued. "AI systems are based on 'training data', which means you need many thousands of examples of content in order to train a system that can detect certain types of text, imagery or video."
So, essentially, until there are significantly more murder, suicide and attack videos captured in Facebook's database, the AI can't be trained to detect and intercept such streams in anything approaching real-time.
And so the reliance comes down to moderators or user reports. "During the entire live broadcast, we did not get a single user report," Facebook reported. "This matters because reports we get while a video is broadcasting live are prioritized for accelerated review. We do this because when a video is still live if there is real-world harm we have a better chance to alert first responders and try to get help on the ground."
As regards the potential to introduce a time delay, as seen in live television broadcasting, the blog dismissed this: "There are millions of Live broadcasts daily, which means a delay would not help address the problem due to the sheer number of videos. More importantly, given the importance of user reports, adding a delay would only further slow down videos getting reported, reviewed and first responders being alerted to provide help on the ground."
So, 'we can't detect the videos and we don't get reports', but 'we can't delay the videos because then we wouldn't get reports'? And 'we can't delay as that would hamper alerts to first responders' even though 'we can't detect the videos and so we can't alert first responders'?
Right, got it.
An expedient sacrifice?
As reported by Fast Company shortly after launch, "Live has been touted as Mark Zuckerberg’s pet project, one he’s 'obsessed' with. Some believe Live is the key to Facebook’s future—a resource that will help it compete against broadcast television."
Three years later, Facebook is focused on riding out the neverending Cambridge Analytica scandal. Reports of criminal investigations, grand juries, undisclosed secret meetings and 'who knew what and when' are debated daily. And as last year was all about the headlines, this year will be all about the investigations. But still, in truth, there's a lack of immediacy and personal impact for the network's 2 billion users. The live broadcast of mass murder has changed that. Politicians and regulators cannot ignore this, they cannot ignore the implications of a platform that live streams material that no modern-day broadcaster would consider, that continues to self-police, that exists in an apparently unregulated bubble.
With the current significant newsflow around the role of social and mainstream media in fueling extremism, the danger for Facebook Live is that it becomes part of the 'concept of operations' for the extremists. The opportunity to live stream shapes or even prompts an attack. Terrorists thirst for publicity. Never in history have they had the opportunity to control the broadcasts.
And so Facebook has a decision to make. A decision that's maybe now more palatable and easily made with the shift in strategy. For now, it's still in their hands, but that could change. The European Union is clearly out to set an example. And when the FTC fine comes down and when U.S. politicians determine how to respond to the latest events, we may see a serious thirst for change in the U.S. as well that will have a more material impact. And if the criminal investigations land badly on Facebook, or if further allegations or breaches come to light, Senator Warren will not be a lone voice calling for very drastic action.
Looking to get ahead of this, to keep up with the rhetoric, if Facebook can't protect everything and win on all fronts, does it sacrifice Facebook Live to protect core data trading revenue streams and throw some weight behind the shift to privacy? They'd obviously rather not, but if the clamor for regulation and even break-up gets more of a tailwind, it could easily become their 'least worst' option.
Over the next few weeks, there will be increasing calls for live streaming services in general and Facebook Live in particular to cease such services. This will gain traction because it’s hard to argue that the public interest benefits from not doing so.
Facebook’s admission that they can’t effectively moderate the scale of content in real time provides little by way of a counter. Right now it will seem hyperbolic to suggest that Facebook Live might fundamentally change or even end. It won’t seem that way a few months down the track.
"I understand that many people don't think Facebook can or would even want to build this kind of privacy-focused platform," Zuckerberg admitted, "because frankly, we don't currently have a strong reputation for building privacy protective services, and we've historically focused on tools for more open sharing. But we've repeatedly shown that we can evolve to build the services that people really want. This is the future I hope we will help bring about."
Okay then, let's see...