What If We Generate Our Own Data? | Rory Byrne

So my name is Rory and my start of Metro is a new way for companies to collect data from their users. It's fast, it's secure, and it gives you access to data from every platform on the internet. So this is a statement that we've all seen before, right? Some of us are even rolling our eyes as we look at us. What I want you to keep in your minds for the next few minutes while I talk about data generation, the problems with the current way that we do it, and then I'll introduce my proposed solution. So right now, all the user data in the world is generated on the corporate side. This is the way that internet companies have been built up until now, and it involves the platform owners generating data about the users in real time. This has a number of problems which I'm going to get into, but the core is that it was built for the advertising era. It's not built for the AI era that we're going into. My proposed solution is a complete inversion of this system whereby the users generate data from all of the platforms that they use. I call it self-generation. So you might notice that I'm focused here on the word generation. That is because, if we're honest, the word collection is extremely misleading. It gives us this impression that we are generating data and that we own it and that we make it available to the companies. It simply doesn't work like that. So I say we ditch the word collect and call it what it really is: data generation. So what do I mean when I say a data generation company? Well, Google was the originator. It started off as an algorithm, the PageRank algorithm. This was Larry Page and Sergey Brin's research at Stanford University, and when they decided to commercialize the research, they turned it into a search engine. And the business model, naturally given it was the 1990s, was advertising. Think about it: this was the golden age of advertising. Not only were people paying for the ads or the products in the ads, people were paying for the cable service to deliver it, and they were also paying for the television that would display it to them. It was the perfect system. So if you think about the way TV advertising actually works, a TV channel wants to monopolize your attention so they can show you ads on the channel. It's very simple and very effective. But in Google's case, it's a little bit more complicated and much more effective. Google does not show you ads directly on their platform, only a little bit. Instead, they have an audience network of over two million websites covering 90% of users on the internet. That's where they show the ads. So why then do they want to monopolize your attention on the platform? Why is Facebook so addictive? We all know this is so they can generate data about us and use that to target ads to us. This is what I mean when I say that the model was created initially for advertising. However, data is useful for a lot more than just advertising. Right now, we're entering a new era with all sorts of useful applications for data in the AI world. So it's not really fair to call Google an advertiser anymore. They're a data generation company. So we often hear statistics like you create a gigabyte of data every day on the internet. That's simply not true. You're not the one creating it. We don't know the who, what, where, when, or why of this data. It's corporate data generated on the corporate side. This has a few problems. First of all, the users have no control over it. They don't know what's generated, they don't know how it's generated. The corporation knows that and they decide what's done with it. Since all the data is owned by the platform, they can use it to construct barriers to entry, stifling innovation across the board. Let's be honest here, guys: this model of traditional data collection is great for the tech giants, it's terrible for everybody except the tech giants. It means that companies have to buy data from data brokers, which is stale, inaccurate, difficult to cross-reference, and it's an ethical minefield. You know, these companies now have multi-industry monopolies and they acquire new companies at a rate of over one per week. That's terrible if you're a small guy with a big idea, deterred. And probably the most striking problem here is that it allows for organizations like Cambridge Analytica and surveillance programs like the PRISM program that Edward Snowden uncovered. It allows these things to be created without our knowledge because all the data in the world is created and used behind the silicon curtain. So I go back to this statement that I mentioned. It should be clear by now: this statement is a complete lie. We're not living in the age of information. The tech giants are. So I propose a solution: we flip the whole thing on its head. Metro is doing that by allowing users to generate their own data on every platform they visit, and then they can exchange that data with businesses and companies on their own terms. This has a few key benefits. Firstly, it's user-centric. It redefines the data collection industry in terms of the users. This is something we've seen before. It's not an innovative idea. Airbnb redefined the hotel industry in terms of the users: anybody can be a hotel. Uber redefined the taxi industry in terms of the users: anybody can be a taxi. Think about the internet: anybody can be a media distributor on the internet. You don't need a van, you don't need newspapers. This is how internet technologies are supposed to look. Secondly, it's mutually beneficial. Businesses get access to gold standard user data in real time from any platform on the internet, provided they abide by a few key laws. The users get to generate the data, the users get to have control. It's transparent. Thirdly, Metro is an open platform, so anybody can build data generation code on top of the platform. What do I mean by that? So Metro projects and companies are powered by what I call data sources. These are open source, community-made plugins that allow users to generate their own data. Currently, Metro supports the major browsers. Users can interactively do things like label images by right-clicking them, or highlight text and enter a translation. They can also possibly track things like their Medium reading habits or your Google search activity or anything else you can think of. So the key benefits of this type of model is that, first of all, the data sources are community-made. That means that anybody, any hacker in their bedroom, can come up with a new innovative type of data and then build a business on top of it. Currently, it's only Facebook engineers can say, 'Hey, what if we got this cool piece of data?' There's no innovation in the space of data generation. It's just not possible right now. Secondly, and possibly more importantly, the person who writes the code to collect data is not necessarily the organization that wants the data. This separation vastly reduces the conflict of interest that is currently causing such a big problem in the traditional model. Secondly, it's open source. This needs no explanation. Transparency is key in this area. Thirdly, it's a one-click install, making this data generation layer of the internet easily accessible to everybody, even my grandmother. And yes, I did actually do user testing with my gran. Whether it's one-click for tests, all these data sources are project independent, which means that a single data source can power multiple projects. First of all, this evens out any cognitive bias that might happen. If you know you're generating data for a certain project, you might act a little differently. If you know you're generating data for an advertiser, the more different projects you contribute to with your data, the more that tends to even out. Secondly, the multiplier effect: each time a user produces a data point, they can be paid multiple times. This is first of all a great incentive for the users to generate data. Second of all, it makes it a lot easier for new companies to get onboard on the platform. If there's already 10,000 users producing data, you just need to convince them to add one new company to their list. So I'll go into a bit of the technical discussion of the actual data source code. I'll go through quite quickly. It's pretty simple. JavaScript object expects two fields: firstly the name, secondly an init function that takes the Metro client as a parameter. This exposes the API. This is the entire code for a data source that allows you to highlight text and enter a translation. So the first thing we want to do is add a right-click menu button. It takes the title of the button and the context in which it appears, i.e., selection. And then in the callback, it receives the context info containing, among other things, the text that we selected. So in this callback function, we want to create a pop-up that the user can enter a translation in. The Metro API has a modal API for you. You just give it a series of inputs to appear in the pop-up form. Currently we just have an input field here. You can add selection fields, drop-down menus, that kind of thing. And then in the callback, you receive all of the data that the user inputted, and it's as simple as just sending it off to the Metro API to be distributed to all the projects. Right now, this is centralized, so all the data comes into Metro and I found it out. The next step is going to be a decentralized data transfer system built with a distributed ledger. So this is what it looks like: you literally highlight your text, right-click, enter your translation, hit enter. This is like less than 30 lines of code. There's a lot you can do with this. So to finish up, I want to highlight the key differences between the current model of data generation and how I think the future will be built. The corporate model is where a corporation generates data about all of its users. This caused a host of problems. You know, it is designed for the advertising era. It sucks for everybody who is not a tech giant. Self-generation means that the users generate the data from all the platforms that they use. This means data can be shared very easily and on fair terms that the user agrees with. This is how the next generation of the internet will be built. It's how it should look. My name is Rory. If you want to get in touch, you see my email there. And the project is live if you want to check out the website. So thank you for listening.