AT THE heart of the internet are monsters with voracious appetites. In bunkers and warehouses around the world, vast arrays of computers run the show, serving up the web – and gorging on our data.
These server farms are the engine rooms of the internet. Operated by some of the world’s most powerful companies, they process photos of our children, emails to our bosses and lovers, and our late-night searches. Such digital shards reveal far more of ourselves than we might like, and they are worth a lot of money. They are not only used to target advertising and sell stuff back to us, but also form the building blocks for a new generation of artificial intelligence that will determine the future of the web.
“Very big and powerful companies own a huge chunk of what happens on the web,” says Andrei Sambra, a developer with the World Wide Web (W3) Consortium at the Massachusetts Institute of Technology, the main standards organisation for the web. But we – the ones producing this valuable data – have lost control.
Advertisement
The time has come to push back. Sambra is part of a growing movement to wrest back control over our digital lives by breaking the monopolies of the server farms and the people who own them. Tweak the technology on which the web runs and we can each keep our own little part of it in our pockets, they say – and determine who or what makes money out of who we are.
In a sense, that would be just getting back to the way the web was always intended. The original World Wide Web, invented by Tim Berners-Lee at the particle physics centre CERN near Geneva in 1999, was a “decentralised” affair. There were no central servers; websites ran on individual machines in universities, offices and bedrooms. Hosting a site just meant plugging a computer into your internet connection and having it serve up the HTML code to anyone visiting. No one company ruled the roost.
Simple open protocols meant that anyone who knew what they were doing could be a part of the burgeoning network. “A lot of the things that made the early web wonderful were these open standards,” says Harry Halpin, also with W3C. “This allowed a level of decentralisation, and lack of monopoly control of the web.”
It sounds utopian, and in many ways it was – but far too fiddly for most people to faff about with. Those open protocols are still there. But we were lured away by convenience.
Hotmail launched in 1996, allowing anyone with an internet connection to have an email account with none of the hassle of running their own mail server. Within a year, 8.5 million people had signed up and it was bought by Microsoft. In 2004, Google launched Gmail, helping us manage our personal lives as well as the web, which it had dominated since the late 1990s.
Facebook also launched in 2004. It made finding and keeping in touch with friends easier and more convenient than earlier social networks like Bebo and Myspace. A decade later, Facebook is used by almost a quarter of the people on the planet. For many, it is all the web they want. It’s where they conduct their social lives, get their news and find entertainment.
Despite its seemingly infinite nature, the web is largely centred on just a handful of companies. Instead of a proliferation of independently run sites, the web is dominated by global firms with which we have made a Faustian pact. In exchange for convenience, we let companies like Google, Facebook, Amazon – and, more recently, start-ups like Uber and Airbnb – conduct their business by siphoning up and profiting from our data.
“The early web was utopian – but far too fiddly for most people to faff about with“
Why should we get worked up about this? After all, the most useful things you can do on the modern web rely on this data. Companies use it constantly to tweak the services they provide. Our data also feeds the machine-learning algorithms that are behind the many recent breakthroughs in artificial intelligence.
“People aren’t too sure how their data is used. This makes them nervous and frustrated“
There are some well-rehearsed objections. One is privacy. Surveys conducted by organisations like Pew Research in Washington DC, for example, which has been studying this issue for 15 years, repeatedly show that people have low levels of confidence in how their personal data is handled by internet companies. “A lot of people aren’t too sure they know what is collected about them and how the data is used,” says Lee Rainie at Pew Research. “This makes people nervous and frustrated, but they still need to live in the modern world and many feel it is not an option to be offline.”
Leaving our digital lives in so few corporate hands also makes us easier to spy on. PRISM, one of many US government surveillance programmes revealed by the whistleblower Edward Snowden, consisted of government agents walking into the premises of large web companies with a secret court order and taking what data they wanted.
A set-up based around servers that store our data makes life ridiculously easy for hackers. No matter how good the defences, the fact that the internet stores its data at actual physical locations makes those computers targets. Servers are honeypots.
Giving ourselves away
Andy Clarke, a philosopher studying artificial intelligence at the University of Edinburgh, UK, says that our loss of control goes even deeper. “When we use the internet in the ways it’s mostly available – through big nodes like Google and Facebook – we are giving ourselves away,” says Clarke. They are making big bucks out of us, and we don’t get a penny. Aral Balkan, founder of Swedish tech democratisation movement Ind.ie, calls such companies “people farmers”. If you’re not paying, you’re the product.
This is worse than you think. The increasing numbers of connected devices in our lives are all sources of data. Soon our data trails won’t begin and end with the time we spend at our screens, they will continue via the smart devices in our homes and offices, on our bodies and in public places. The artificial intelligences being created by internet companies will make us ever more dependent on their services. Coupled with this is the rise of decision-making software, which firms are increasingly using to help make calls about loans, job applications and health insurance based on your data.
“The public really does need to be aware of what this means,” says Mohamed Sayed, CEO of German start-up Heuro Labs, which develops health AIs. Our data is being used to train these systems, he says. “How does society get back some of that?”
It’s going to require a radical rethink. That’s just what Sambra and his colleagues at W3C are pushing for. The aim is to combine the control and personal autonomy of the early web with the ease and usefulness of the one we have today.
Sambra is working on a project called Solid, which is led by none other than Berners-Lee himself. The idea behind this prototype software is to separate our data from the apps and servers that process it. With Solid, you get to decide where your data lives – on your phone, a server at work, or with a cloud provider, as it probably does now. You can even nominate friends to look after it. “We want to put the data in a place where the user controls it,” says Sambra.
Instead of sucking up your data by default, apps will first have to get your permission. And rather than having its own database to draw on, an app will pull in data from as many sources as it needs. In Solid’s vision, web email and social networking sites would cease to handle personal data and focus instead on building the type of flexible software that can draw on data from anywhere.
In particular, Solid aims to give you ownership of one of your most important pieces of data: the list of the people you know, and the people they know – what’s called our social graph. A big reason we get locked in to certain sites and services is that it is hard – often impossible – to take this social graph with you if you leave. Quit Facebook, Twitter or LinkedIn and all of your connections and contacts are lost. With Solid, you carry your social graph with you from site to site, freeing you from becoming locked in. “You have one social graph and you can reuse it in any app,” says Sambra. You control which parts to plug into which networks.
In Solid, your social graph is a key personal possession, a digital Rolodex not to be handed over to anyone. It’s important not just as a list of who you know, but as a list of who you trust. This comes into play when deciding how to distribute your data. You could carry this social graph in your phone, says Sambra. But it is impractical to have crucial pieces of data stored only in one place. So, like leaving sets of house keys with neighbours, you can choose to entrust sensitive data to close connections in your social graph. Web services will then need to adapt to such fluid arrangements.
Solid is still at an early stage. But it is not alone. This year, UK start-up MaidSafe launched a peer-to-peer network that relies on encryption and the blockchain – the distributed ledger technology that underpins bitcoin – to divorce data from servers. It already has a few thousand users.
MaidSafe’s approach is pretty radical. Where Solid would operate as a virtual layer on top of the existing structure of the internet, MaidSafe’s network does away with servers completely. Instead, it asks everyone who joins to contribute a little computing power and storage (see graphic). To join, you simply download their software, and it is this, rather than central operators like Facebook, that encrypts your data and keeps track of it.
A big advantage of this approach is with no servers, there are no targets for attackers. No system today provides physical security for your private data, says MaidSafe’s founder David Irvine.
Getting rid of servers throws up some very basic questions, however. For example, how do you log in to a website if there’s no server at the other end to deal with the request? Irvine’s answer is to log into the network itself – which consists of whatever computers happen to be online at the time.
For contributing to the running of this serverless internet, users earn a bitcoin-like cryptocurrency called Safecoin. This can be exchanged for services on the network or converted to other currencies. MaidSafe’s fledgling community has already developed a handful of apps, including a blogging platform, a file-sharing application and a basic social network. Email and video conferencing are in the works.
Solid and MaidSafe are far from the first attempts to gain some control over the way we use the web. The social network Diaspora lets you keep hold of your data, and so-called personal clouds provide online storage but let you keep your files and photos on a hard drive at home. Yet nothing quite matches the convenience of the web we have.
The difference this time is that the slick interfaces of smartphones and web apps are much better at hiding the technical details. Things just work. Better technology means that from our point of view, the way we use the web won’t appear to change. We can have convenience and decentralisation.
“We have tech that Tim Berners-Lee didn’t have 25 years ago,” says Brewster Kahle of the Internet Archive in San Francisco. “There’s a real shift going on. The internet’s plumbing is a distributed system, the World Wide Web is not. I think we could make that jump.”
Indeed, decentralisation is in the air. Drawn by the promice of a more reliable and more secure internet, IBM, Samsung and the US military’s research arm, DARPA, are all working on their own versions of such networks. But building a distributed web that better serves those who use it is hard. It won’t happen soon. The internet companies that control the web spend billions of dollars a year on the infrastructure and engineers needed to keep their centralised systems running. Folding that all into something that can run across many machines controlled by you and the people you trust is not just a massive technical challenge – it’s a social and political one too.
“We have tech that Tim Berners-Lee didn’t have 25 years ago. There’s a real shift going on“
The answer, says Halpin, is for the developers working on different parts of the distributed web to start talking to each other about their work, something that doesn’t currently happen. “The community has to get together with the adequate expertise and solve these hard problems and push open standards,” he says. Open standards will make it easier for talented developers to build applications without having to go through existing networks.
We might even find ways to take back a tiny stake in the fortunes being made. Sayed sees a system in which we retain ownership of the small pieces of data generated as we move through this sensor-laden world. Whenever our data is used in a way that makes revenue for someone else, we would get a tiny fraction of that money, a small reward for our part in creating value. The same scenario is also championed by tech guru Jaron Lanier in his 2013 book Who owns the future?. “There is an emerging vision that if people could control their data themselves, not just a few companies, it would have a massive positive impact on society,” says Halpin.
The ways of the web are entrenched. Is it too late to change them? It’s hard to say for sure. The incumbents will take some pressure before they budge. The result may be that they adopt their own version of Solid or MaidSafe. But that would still be good for most of us.
It’s also worth remembering that the internet is only a few decades old – still young enough to learn new tricks. Until recently we simply hadn’t any viable alternative to how things are done. That’s about to change – if we want it to.
This article appeared in print under the headline “The web we want”
