DO YOU remember your first Google search, with the world’s information at your fingertips? Or perhaps the thrill of rediscovering old friends when you opened a Facebook account. From those heady, early days, Google and Facebook have become custodians of crucial technologies relied upon by hundreds of millions of people all over the planet.
The key to their success is no secret: the services they offer are the best and so droves of people choose to use them. But with the droves comes data and with that, a new form of power. This, coupled with the fact that most users understand little about this new currency, has led some to worry that there is very little holding these companies to account.
“They really control what you know and what you see. They control your universe,” says Ariel Ezrachi, who studies competition law at the University of Oxford. Is it time to rein them in? Even if we wanted to, could we?
Advertisement
Few would argue that the existence of Facebook, Google and other information mega-companies has been anything other than a boon. But it’s not hard to imagine scenarios that raise concerns. Facebook could tip the balance of the US presidential election, say, with a small interface tweak that preferentially nudges supporters of one candidate to vote. That’s not too far-fetched: the firm is thought to have increased voter turnout in the 2010 US congressional elections by , essentially peer-pressuring their friends to do the same.
Facebook is increasingly people’s gateway for news: in 2016, , up from 31 per cent in 2013. Since its news feed algorithms control which stories people see, there are concerns that the company has the potential to by curating the news.
“Google and Facebook control what you know and what you see. They control your universe“
Similar scenarios could be imagined with Google. It could start charging insurance firms to access group insights from its massive user base, say, which could mean individual customers’ premiums go up. Or it could close down services that millions of people find useful because they don’t bring in the right kind of data or adverts to make money.
At the heart of these companies’ power is user-generated data, tonnes of it. It lets them make huge amounts of money from firms that want to use targeted adverts, and also to cut deals with a host of entities – public and private – hungry for the knowledge that can be harvested from massive data sets (see “The small print“).
We are looking at a future where information companies become the enablers of services in all domains of life. “My prediction is that we’ll look back in 10, 15 years on this period as a remarkably naive and irresponsible time,” says Julia Powles, a researcher in law and computer science at the University of Cambridge.
Consolidating their power is the fact that while the companies understand the value of this data, the people they collect it from have little idea of its worth, now or in the future. “Neither the regulators nor society understand the real value,” says Ezrachi.
This knowledge gap explains the recent handovers of data from the UK’s National Health Service to Google’s artificial intelligence arm, DeepMind, says Ezrachi. DeepMind understands the value it can extract from millions of patient records – in this case retinal scans – while just the chance that DeepMind will reduce the burden of eye disease is enough for the NHS. Giving DeepMind access to that data for free may not have been the best option for the UK public, who ultimately own the data, but there was no discussion about the deal beforehand.
A market of one
Facebook and Google now look like natural monopolies, says John Mark Newman, a legal scholar at the University of Memphis, Tennessee. This makes sense for the consumer – if the purpose of a social network is to connect people, it’s easier if everyone is on the same one. But natural monopolies should come with accountability.
In the physical world, public utilities like gas and electricity in the US are assigned regulating bodies known as public utility commissions, answerable to an elected official, to ensure they don’t abuse their power as a natural monopoly. Networks built and controlled by information companies have no such oversight. “What we have is a system that relies on common sense and the good will of those at the head of those organisations,” says Ezrachi. “That’s it.”
Public utility regulation is usually done by overseeing pricing. “All we did was regulate prices with electricity, gas and plumbing,” says Newman. “That really was a rule of thumb that worked pretty well for 70 or 80 years.” But this won’t work for Google and Facebook for a simple reason: their products don’t cost us anything – at least, not cash.
Other traditional responses to monopolies don’t feel right either. The idea of “breaking up” Facebook seems ridiculous as its entire usefulness comes from having everyone on one network. The same goes for Google, Microsoft and Amazon. Each is useful because of its scale.
So what can we do to ensure these companies stay on the right side of benign? “This is a massive question, there isn’t an easy answer,” says Ezrachi. “Any approach has to be mindful of the risk of chilling innovation. On one hand you have extreme benefits – you have a lot of good stuff that comes out of these technologies. On the other, you have gatekeepers that are able to occupy all the critical junctions.”
Although firms like Google and Facebook can’t be controlled by regulating pricing, they could still benefit from independent oversight. A board could examine data collection and processing practices, says Newman. DeepMind has taken a step in this direction by appointing an independent review board to oversee the activities of its healthcare division. It had its first meeting in July, which focused on procedural issues.
A priority should be recruiting technically minded people into existing regulatory bodies, says Powles – people who understand how large technology platforms are affecting individual autonomy.
“There is no regulation. What we have is a system that relies on common sense and good will“
Another option is to enforce existing laws with more enthusiasm, says Richard French, legal director at Digital Catapult, a UK-based non-profit. New European Union regulations called the General Data Protection Regulation, which will apply from 2018, may help as companies will be subject to far larger fines for data protection breaches than currently exist – up to 4 per cent of global annual income in the more serious cases, says French. “That should really focus data controller’s minds.”
The catch-all fix is education, helping consumers understand the value of their personal data, and why it matters who controls it. This is particularly important in the age of machine learning, when our data can be used to train software that can perform autonomous labour.
For example, from next year, all cars sold by Toyota in the US will come equipped with sensors that enable some autonomous driving features. But they will also collect data that will become the bedrock of Toyota’s future software for driverless cars, potentially securing its growth for decades to come. If consumers don’t understand that this is happening, then there can be no informed discussion about whether it’s a good idea, or if and how consumers should be remunerated for it. “Sometimes what is free in the short term would cost you much more in the long term,” says Ezrachi.
Our information giants are barely a decade old – what happens in coming decades is anyone’s guess. The history of the automobile offers a useful parallel. While the autonomy of movement that cars provide has transformed the world, 1.25 million people are killed each year in accidents. That number would be a lot higher if Volvo hadn’t invented the seat belt. Ezrachi sees a parallel: “I’m not saying stop driving cars, but let’s have a few seat belts.”
The small print
Digital services have become a large part of our lives, but they come at a cost. Here’s what we give away when we use them.
Instagram As well as being a huge audience to serve ads to, users of the photo-sharing app are giving parent company Facebook hashtags, which its uses to train machine learning systems that handle images.
Facebook The words you type and the clicks you make are used to teach machine learning systems what you are interested in, so it can show this in your news feed. Your on-site conversations are anonymised and used to train its systems how to hold human-like conversations.
Google search After analysing billions of searches, Google knows that humans aren’t as unique as we think. It’s algorithm has learned so much about what we want that the site has started to display the clearest answers right at the top of the search results, above any web links, dictating how we find out about the world.
WhatsApp Last week, the messaging service announced that it is sharing some users’ data – including phone numbers – with parent company Facebook, opening another way for business to reach customers. The move provoked an angry reaction with many people saying they will no longer use the service.
This article appeared in print under the headline “Taming the tech giants”
Article amended on 1 September 2016
The date for compliance with the General Data Protection Regulation has been corrected
