While the coronavirus crisis has been distinctly not great for humans, it’s already been a boon for robots. As humans have stayed home to slow the spread of the disease, robots have entered our lives in novel ways, from delivering groceries to taking over manufacturing jobs that put humans at risk of infection.
But this shift, which came faster than some experts anticipated, comes with concerns over individual privacy and security.
Quartz spoke with PW Singer, a security expert and one of the authors of the new book Burn-In, a techno-thriller that leans heavily on reality. The authors’ last book, Ghost Fleet, won Singer and his co-writer August Cole invitations to the White House, Congress, and the Pentagon to discuss its real-world impacts.
Singer, who is a strategist at the New America Foundation, was one of the experts who warned of the Covid-19 pandemic early on and offered predictions of its long-term impacts. Though he and Cole wrote the book over the course of the past few years, they’ve seen signs that the pandemic is hastening some of the global changes they uncovered in their research, including greater automation and the deep privacy concerns that come with it.
This conversation has been edited for clarity and length.
What, exactly, are some things you’re seeing accelerate due to coronavirus?
Many of the trends that Burn-In explains were in place before the pandemic, and every data point is illustrating how they’re coming even faster now. We have one generation that is doing remote learning, another that is doing remote work at levels never expected. We have fields like telemedicine that in a matter of weeks jumped ahead to where we thought we would be in 10 years. Robotics are being deployed into roles like policing curfews, to cleaning subways and hospitals, to delivering groceries.
And finally, you have mass surveillance via AI of individuals and society scaling, well past what even some of the most extreme science fiction imagined.
After this pandemic, we’re not reverting back to the way things were before. That also means that all the tough political, economic, security, social, and even family issues that our characters wrestle with, we’re going to face them in the real world sooner.
How are mass surveillance and AI going past what sci-fi imagined?
You are seeing both an acceleration and an extension of the amount of data collection on individuals and society, including extending into areas that weren’t even being thought about just a few months back. It’s not just people’s buying habits or their communication—it’s integrating everything from past medical records to tracking body temperature. What happens next is that AI takes that tracking and moves it into prediction of behavior and then into influence, shaping that behavior.
Take facial recognition software. It’s not just matching your face to an identity and then that identity’s digital history. It’s also taking that data and making projections for what happens next, where that person will move, buy, or vote next. The key is influence—how to shape what the individual does next.
For example, in the book, a character is walking through Union Station in Washington, DC and is identified via facial recognition, and then ads are popping up that are seemingly tailored to them, seeking to divert them. But then there are ways that the individual could be influenced and they don’t even know it’s happening to them. A recommended route that is subtly shifted, a series of posts that pop up in your feed that changes your worldview.
That’s not futuristic. There is even a company you can hire that will subtly insert articles into your significant other’s Facebook feed, designed to get them to change their mind on something, such as to influence them to pop the question or to convince a spouse to get a pet.
How does that work if no one is going anywhere?
One of the few economic sectors that has boomed during the pandemic is robotics. They’re not broadcasting it loudly, but the companies are doing super well because demand has spiked.
Robotics can work in a mix of roles, and companies that range from factories to banks are thinking into the next few years about how they can keep from having to shut back down in the next wave of Covid-19. For an assembly line or cleaning robot, it’s not just that the system doesn’t have to be paid a wage, it doesn’t get sick, it doesn’t have to shut down the next time there’s a social quarantine.
So what do we do now?
This wave of AI and automation is not just the most important tech around us right now, it might be the most important in all of human history. It’s not just the equivalent of the last industrial revolution, but a new kind in that for the first time, it is not just swapping the tool—a farmer’s shovel for a factory worker’s hammer—but making the tool intelligent and thus affecting the worker.
And yet we have a disconnect. Ninety-one percent of leaders think AI is the most important piece of technology out there. It’s woven into the strategy plans of everything from the US military’s new National Defense Strategy, to the Chinese government’s strategy of dominating the future, from big tech companies like Google and Facebook, to John Deere and McDonald’s, both of which just bought up AI companies because they think it’s important to their future. But only 17% of leaders self-reported that they had familiarity with AI, let alone an understanding of its implications.
Think about that delta between 91% and 17%. We have to go after that explainability problem. We have to understand AI. I don’t mean that we have to know how to program it, I mean we have to understand the basics of it, the key terms, the key issues and dilemmas we’re going to face. And we’re not going to face them in 50 to 100 years, like the Secretary of Treasury wrongly said—it’s happening right now. Eighty-five percent of manufacturing job loss of the last generation was to automation, not to jobs going to China.
What could slow that job loss?
A key issue moving forward is human-machine teaming. You’re definitely getting these robots, there’s no stopping it. So what are the roles that are appropriate for machines in settings like the military, hospitals, banks, and media, and how will humans work with the machines?
From there we have to consider how we train not just the machine, but also the people working with it. That could affect everything from changing our education system, so that we’re not teaching kids in ways that are going to fail them in the future, to helping people in any industry to thrive in this world.
In turn, there will be areas where we decide that automation shouldn’t go, or has gone too far. We should talk and decide those rather than allowing tech companies to beta test on society. In short, we have to stop thinking of this all as science fiction and instead begin to understand the real-world applications and dilemmas of AI that we’re going to face. From that, we can go after the problems to steer them towards the more utopian view.