3. Yes sir, we have barbed wire going up.

"We're in the age of surveillance capitalism. It's all about the data."
    Liz Bristo looked out at her students. The university lecture theater had been built in the fifties. It had stuffy airflow, bad acoustics, and potentially asbestos in the walls. They had finally attached wifi routers to the ceiling panels a few years ago; one of them looked like it had lost a screw and was in danger of coming down on the unsuspecting students below. Not that half of them would notice over the oversaturated glare of their cellphones.
    "Connect the dots. All of Silicon Valley runs on venture capital. These guys - and they're mostly guys, by the way - put money in these small, risky companies, and then they need them to grow like hell so they'll get thirty-seven times their money back. What characteristic of a normal business do you think makes growth slower than a venture capitalist would like it to be?"
    Liz looked out at the students. A few hands went up from the half of the room that was paying attention; she picked one. "Sonia."
    "Is it work-life balance?"
    Liz laughed. "Work-life balance isn't what I'm thinking of, but yeah, a lot of these companies have really tough working hours. Even though companies started by people in their forties and fifties do better, people like to fund startups that are run by these 23-year-old prodigies, because they don't have families, and they're not smart enough to know better than to work eighteen hours a day, seven days a week. A lot of of these companies provide two or three catered meals a day, and stuff like laundry services, so nobody ever has to leave the office. That was a great suggestion. What else?"
    Fewer hands. Liz picked another one. "Carrie."
    "Is it making money?"
    "That's the one I was thinking of. If you try and actually make money, you can't grow as fast. If people actually have to get out their credit cards and pay something to join a service, they're going to be much less likely to go ahead and do that. And suddenly they're not going to get the seven percent week on week growth they're looking for. And there's another really nasty side effect that these entrepreneurs try and avoid. Does anyone know what it is?"
    No hands.
    "I'll tell you. If you're actually bringing in revenue, the value of your company is tethered to the amount of money you're bringing in. You might be able to argue that you're worth seven times your annual revenue, or even ten times. But you're never going to be able to argue that you're worth a hundred thousand times your revenue. On the other hand, if you're not making any money at all - then you can make all kinds of crazy arguments about what you're worth and there's nothing to really say that you aren't worth. I mean, you might be, right? But you also might not be.
    "So what these companies like to do is grow really fast, get a lot of users and, most of all, their data, and then switch one some kind of revenue stream once they're sure they're going to make enough money to justify a high enough valuation. But even then, they need to keep that growth rate up, so they can promise new investors that they're going to be worth exponentially more in the future. So unless they're targeting large enterprise businesses who are going to pay them a ton of money per customer, they still can't just take money from people. So what do you think the solution is?"
    A few more hands. "Jiyeon."
    "Advertising?"
    "Yes! Advertising. Pretty early on, internet companies realized they could put display ads on their site and make money based on how many people were seeing them. They used to talk about eyeballs: a site might be getting a million eyeballs a day. In the original dotcom bubble 18 years ago, you could just put up a banner ad on your site and that would be enough to pay the Lizs. In fact, some people think that the banner ad on the Yahoo homepage fed the whole dotcom bubble."
    She stopped for a second. "Wait. How many of you know what Yahoo was?"
    Even the students who were glued to their phones looked up for a moment. Most people in the room - but not everyone - raised their hands.
    "Okay, so Yahoo was what we used to call a 'portal'. You could search for websites on it, but it was meant to be the first thing you saw when you visited the web. It had news, links to new websites, a little widget that gave you a weather report, and so on. And it had this banner ad at the top of the page, just four hundred and sixty-eight by sixty pixels, which would advertise brands with a static image or an animated GIF.
    "And what would happen is that new startups would advertise on the Yahoo homepage banner ad. They'd suddenly get a ton of traffic, because everybody used Yahoo as their homepage, which would increase their perceived value. They would use that value to raise more money from investors, which they would then use to buy more ads on the Yahoo homepage. Meanwhile, Yahoo became incredibly valuable because it was making money from all these startups. More investors came in because everybody seemed to be making all this money. But it all turned out to be this insular, terrible machine, revolving around a single banner ad on a single homepage. And when the house of cards came down, it brought down the tech industry and dragged the global economy with it.
    "So after the dotcom crash, a banner ad wasn't going to cut it any more. The prices went through the floor. But tech companies were already addicted to advertising - so they needed to find ways to make it more valuable. And it was actually an advertising company called eUniverse that saw the potential of social networks to create vast, dedicated audiences that you could sell ads to, and created a site called MySpace.
    "Meanwhile, Google - which hadn't figured out how to make money yet, but had grown really quickly - saw the potential of taking peoples' information and selling advertising that was precisely targeted to them. They used their investment money to buy a startup out of Santa Monica called Applied Semantics, which had some technology to make targeted advertising better.
    "And suddenly you had an arms race. Tech companies needed to gather more, better information about their users in order to sell those ads so they could pay their Lizs and become valuable companies without stunting that fast, hockey-stick growth that made them attractive to investors. So for the next fifteen years, most of the most advanced technologists in the world went to work for companies that were trying to figure out how to spy on their users."
    A hand went up. "Yes. Sonia."
    "Why did they need to spy on their users? I have a Facebook profile, for example. I added some information about myself. Isn't that enough?"
    Liz laughed. "You would think, but absolutely not. It turns out that the information you supply voluntarily is nowhere near as valuable as the information that can be figured out about you based on your activity.
    "I'll give you a real-world example outside the technology industry. Target, the big box store, has a huge treasure trove of information about what you've bought over time. Even if you don't have a rewards card account with them, they can figure out who you are based on your credit card details. They can link that to other databases and figure out trends across all their customers. So, for example, they know when a customer is pregnant. It's doubtful that a customer will just say, 'hey, Target, by the way, I'm pregnant', but they can work it out from the changes in what they buy. Obviously something like a crib or baby toys is a dead giveaway, but well before that, it turns out that pregnant women start to buy different kinds of lotion and things like that. And if Target can figure out that someone is pregnant before anyone else, they can send them coupons for that bigger stuff, those cribs and toys and what-have-you, and win that business. Target and all those other stores were doing this well before machine learning and artificial intelligence.
    "Now that we have AI, it's even more precise. Tech companies gather as much data as they can about you, and machine learning algorithms can identify trends in that data automatically. They create these incredibly micro-targeted categories of people that advertisers can pay to access. And that's much more valuable than a plain old banner ad, because advertisers can be much more certain that a potential customer is going to be interested in their product. And what's happened is, indeed, that the most valuable targeted ads have created incredibly valuable companies, who are making a lot of money through them. It's taken over the technology industry. And it's changing whole cities. San Francisco is an incredibly expensive place to live - a one bedroom apartment costs thirty-five hundred dollars a month to rent, if you can believe it - because of all these people who are getting incredibly high paychecks based on these targeted advertising dollars."
    Liz paused. "Of course, there are a few problems with that. Beyond the housing crisis in San Francisco, I mean. What do you think some of those might be?"
    A single hand. "Harry."
    "The onus is now on every technology company to gather as much information as they can about everyone who uses the internet."
    "Yes. Exactly. Take Google, which has some code running on nine out of ten websites on the internet. They gave away website analytics software, so anyone with a website could, for free, learn more about their visitors. But of course, in the process, Google learns more about their visitors too. And they're able to follow people from site to site to site, learning what they buy, what they're interested in reading, what they watch, and so on. It all feeds their machine learning algorithm. They bought that software, by the way; they acquired it because it fit in with this business objective of gathering all this data. These companies all connect their user accounts to databases run by the credit card companies, so they can link your activity online with your purchasing activity in the real world.
    "And then another thing they bought in was something called Android, which now powers most smartphones in the world. These phones know where you are geographically, who you're talking to, when you wake up, when you go to sleep, who you meet up with, and so on. Google Maps is able to tell drivers with incredible accuracy that there's traffic up ahead because all the cars that are actually stuck in the traffic jam have phones in them that are running Google Maps. So on top of knowing what you read and knowing what you buy, these companies all know where you are, who you associate with, and what you do out in the real world. It's surveillance on a massive scale that totalitarian governments could only dream of.
    "And speaking of totalitarian governments, what else do you think this data collection might make possible?"
    Everyone's hand shot up. Liz smiled. "Okay, this makes me happy. I'm just going to close my eyes and point at someone at random." She closed her eyes, spun herself round, pointed randomly into the audience, and opened her eyes again. "Hi. Rachel. Go for it."
    "If you know all this information about people, you know how they lean politically, and you know exactly what kind of message will work best on them if you want to swing their vote."
    "Bingo," Liz said. "These kinds of ads aren't just incredibly valuable to people who want to sell something. Billions of people use the internet as their source of knowledge, the way they learn about the world, and where they get their news. The average smartphone user picks up their phone up to a hundred and fifty times a day. The average active Facebook user looks at their feed fourteen times a day. It's not just that the internet is woven into the fabric of modern life; the internet is the fabric of modern life. And the tools on these platforms, enabled by venture capital, surveillance technology, and machine learning, are gathering as much information as they can about us. So when a politician really wants to win an election, they just have to use these platforms and run the best information warfare campaign.
    "In the 2016 US election, Facebook actually had a team embedded with the Trump campaign, helping them to to use these targeted ads as effectively as possible. You might have heard of a company called Cambridge Analytica, which used the Facebook API to read through peoples' profiles and interests in order to make predictions about how they could be affected with messaging. That wasn't abuse or a hack; it was how Facebook was designed to be used. And of course, we all know by now that foreign entities were buying ads to try and influence the election, too, because what happens in the US has such a deep effect on politics and commerce across the whole world.
    "But it gets worse than that. In Myanmar, communities that just had a single percent higher Facebook use were significantly more likely to experience genocidal violence. Targeted advertising also had a significant influence on the election in Brazil, where they just elected a literal fascist into power.
    "And it's gone completely unchecked, because the technology industry is horrible at self-policing, and because it in itself is responsible for so much economic value in the US. Politicians are starting to wake up, and there's legislation like the General Data Privacy Regulation in the European Union that is trying to curtail some of the worst behavior, but it's all been too little, too late. Even privacy activists tend to focus on tools that just help individuals to keep their data secure, when we need to think of privacy like a vaccine: it only really works if we have herd immunity. I can be as private as I want, but if my friend has all my details in their phone and they don't take care of their own privacy, all my efforts are for naught. In short, we need to change the way the internet works, and turn it into something that is by the people, for the people."
    Liz looked at the room full of students and smiled. "And that's where you all come in."