David Karpf, associate professor in the George Washington University’s School of Media and Public Affairs (housed in the Columbian College of Arts and Sciences), is an expert on political communication. His book, “Vaporware, Inc.: How Silicon Valley Sells the Future to Control the Present,” will be published by Norton in January. We asked him to tell GW Today about what he calls “techno-pragmatism” and why more regulation of artificial intelligence (AI) is needed.
For the past 30 years we’ve been listening to the stories told by Silicon Valley. They promise that soon we will have this glorious future where there is unlimited wealth and also they have cured cancer and maybe we all live for forever. And that’s always something like 10 years in the future, and it never quite arrives. But in the meantime, Silicon Valley gains more wealth, more power, fewer limitations. We’ve been listening to these stories and trusting them and never noticing that they don’t come out the way that we expect.
One of the traditions in Silicon Valley is that startups are supposed to make big promises that they can’t accomplish yet. And once they’ve found a market for their product, then they’ll go ahead and build it. They said, ‘We’re building the future; you should judge us on our ambitions, not on our results.’ When Silicon Valley was still tiny, maybe there was something to that idea, but these are multi-trillion-dollar companies that the entire global economy is built around. We should judge them based on results, not ambitions. They don’t like that, and that's an awful lot of what led to this reactionary turn in Silicon Valley—this sense that they are not supposed to be regulated, that they are supposed to be special and different.
The ideological project of Silicon Valley over the last 30-plus years is what they call techno-optimism. A major venture capitalist, Marc Andreessen, wrote a piece in 2023 called “The Techno-Optimist Manifesto.” It was a terrible piece of writing, but what he and others are saying, essentially, is that the engineers, entrepreneurs and investors of Silicon Valley are the ones who create the future, who build tomorrow, and the role for the government and for journalists and the public is to applaud and stay out of the way.
I end my new book, “Vaporware, Inc.” with what I call a “Techno-Pragmatist Manifesto,” arguing for a different approach to all of this. To take deepfake voices as one example, the solution is regulations and assigning financial blame. We can’t put the genie back in the bottle in terms of the technological capacities here, but we can offer huge fines to tech firms. To give another example, as soon as they came out with DALL-E, the generative AI image generator, observers predicted that it could be used to make fake porn and harass women in very obvious ways. And of course, that’s exactly what happened.
That doesn’t mean any of this is inevitable. It means that if we want the internet not to be full of harassment of women, then we’re going to need government regulators who say that if a company allows that, they will be fined a lot of money and or given jail time, and then the companies will figure out how to prevent that. But Silicon Valley is being treated as special, as though the rules don’t apply to them. And recently, they kind of bought the government, so now they do whatever they want.
Europe is better at regulating than us right now. We as a country need to step up and reclaim the capacity to actually make demands on Silicon Valley.
AI isn't going to disappear, but the type of AI future that we end up living in depends on regulatory decisions that get made today. It depends on how citizens, journalists and governments make demands upon Silicon Valley to adjust the course of the future. It is through regulation that we influence what things will be profitable and what will not. The demand from the AI industry has been, ‘Let us self-regulate, let us make all of our own decisions and the best futures will occur.’ That’s not going to work out well for society.
They’re always projecting a future that explains to investors why their early investment in AI will be totally justified because there will be so much money to be made in this stuff. They’re telling us the same stories they always tell in order to avoid regulation and in order to get money from us right now. And if history is any guide, the promises they’re making are not going to hold true. And when they don’t come true, they won’t have to give the money back.
Sam Altman, the CEO of OpenAI, wrote a piece called “Moore’s Law for Everything,” a manifesto on how the AI revolution is going to change the world—how law operates, how medicine operates, how people get educated. Mark Zuckerberg recently put out a 6,500-word manifesto mentioning how education’s going to be radically different soon because AI is like a perfect tutor in your hands. They are now so divorced from the life and world of everyday people that many of the things they promise seem like a dystopia.
Since 2020, Karpf has taught a class called The History of the Digital Future. His previous books include “The MoveOn Effect” (Oxford University Press, 2012), about how groups like MoveOn use the internet to be more effective at politics, and “Analytic Activism” (Oxford University Press, 2016), a further examination of how the digital sphere has transformed political advocacy.
