How can physics help solve messy, real world problems? How can we embrace the possibilities of AI while limiting existential risk and abuse by bad actors?
Neil Johnson is a physics professor at George Washington University. His new initiative in Complexity and Data Science at the Dynamic Online Networks Lab combines cross-disciplinary fundamental research with data science to attack complex real-world problems. His research interests lie in the broad area of Complex Systems and ‘many-body’ out-of-equilibrium systems of collections of objects, ranging from crowds of particles to crowds of people and from environments as distinct as quantum information processing in nanostructures to the online world of collective behavior on social media.
NEIL JOHNSON
It gets back to this core question. I just wish I was a young scientist going into this because that's the question to answer: Why AI comes out with what it does. That's the burning question. It's like it's bigger than the origin of the universe to me as a scientist, and here's the reason why. The origin of the universe, it happened. That's why we're here. It's almost like a historical question asking why it happened. The AI future is not a historical question. It's a now and future question.
THE CREATIVE PROCESS
Before we get into your fascinating work at DON, the Dynamic Online Networks Lab, tell us a little bit about your upbringing and educational journey that made you the physicists you are today. I guess some young people growing up wouldn't say, “I want to study data science and solve complex real world problems when I grow up,” but were you like that?
JOHNSON
I was never very good at things like crosswords and math puzzles. I was really more interested in messy real world problems. I remember listening to the news a lot. But fast forward, I start taking classes in school, and end up in physics. I happen to do a natural science undergraduate degree at Cambridge, which did feature physics prominently. It was built around the idea that in the world that we want to understand, just one discipline isn't enough. And yet, I fell into doing a PhD in physics in Harvard. But I was never really happy with that. I remember printing out my PhD and thinking, “that was a waste of time,” because I now know probably more about that tiny little project than anyone else in the world, and who cares? I was reading about something called a Santa Fe Institute for Complex Systems as I was graduating. I became a research fellow in Cambridge for a few years and then a faculty member in physics at Oxford University. During that process, I started exploring other projects. I've always been interested in the idea that messy problems might have some kind of solution if you could get data on them. That launched me off into looking at all sorts of other disciplines.
THE CREATIVE PROCESS
It's intriguing to me to think of people as particles. There's a lot of variance, but if you draw far enough back, they can be like particles. So I imagine your work is a lot like translation work, translating figures from one field to all these different disciplines.
NEIL JOHNSON
You are absolutely right. I learned all this physics, but actually I’m more interested in other types of problems than something sitting at minus 273 degrees that I'll never even be able to see. The typical criticism of physics is that they treat objects as identical. Nobody can explain why the Titanic went down one water molecule at a time. All of those factors went together, none of that explanation comes from a simple assumption of particles being identical and all. So the idea that there's heterogeneity, which is the core idea of all the social sciences, the idea that you could bring that into physics is actually new. All the assumptions physicists make is because particles can be treated as identical, but doing so brings you to conclusions that don't correspond to the messy real world. So in our projects, part of the creative process is learning how to interact with other disciplines. It’s not something academia in general does. I happen to be a saxophone player. In something like music, the richness comes from different instruments—their shortcomings and advantages combined together. My hope is that's what we end up doing in the academic field to try and address some of these messy problems.
THE CREATIVE PROCESS
Your work is mostly funded by the U.S. Department of Defense. We have certain ideals to how a functioning democracy should work. I understand you're trying to be a safety filter to help protect society from bad actors. I’ve noticed in your research of online extremist groups, you've drawn parallels between their behavior and that of natural systems such as fish schools.
JOHNSON
A lot of our work is comparative. We look at background behavior. Is there a burst of new activity? We zoom in on that and ask why that is suddenly appearing and why it didn't appear before. Imagine one day you wake up and you find water in a pot is boiling and you want to understand why water is boiling. If you go at it one molecule at a time, it's not giving you the big picture of what is going on. We've probably all done this: you take milk, stick it in the fridge, too lazy to go to the grocery, so you just leave it there. The 11th day, the milk's gone bad. Why did that happen on the 11th day? What was happening was that all you could see was the kind of macro level, you couldn't see the individual pieces of milk. This is a new area of physics, exactly the same as how shock waves—a wave that builds up so quickly, there's no kind of precursor—appear. Using the data we collect online, we have a tool for making predictions of when we expect shocks to arise and what shape they'll have. So the reason we went for a systems level view is because you can't understand water boiling one molecule at a time, you can't understand why milk produces that curd on the 11th day, one molecule of milk at a time. And I’m no expert in animal societies, but it turns out when we look at the data, online communities that promote hate and extremism scatter when some threat comes—what an anthropologist would call fusion and fission.
THE CREATIVE PROCESS
Yes. All these things that are developing slowly are subterranean, in the dark and we don't know and it's accumulating. You spoke about how AI can be used to accelerate and propagate the labor of writing hate speech. Bill Gates said that AI will change jobs for everyone except those who create AI. So with the change of many traditional jobs, and people perhaps not finding purpose and meaning in their life because their jobs are under threat, there is a growing sense of insecurity that feeds into a need to find communities online, sometimes among extremist groups.
JOHNSON
I'm a huge optimist for AI, actually. I see it as part of that process of climbing its own mountain. It could do wonders for so many areas of science, medicine. When the car came out, the car initially is a disaster. But you fast forward, and it was the key to so many advances in society. I think it's exactly the same as AI. The big challenge is to understand why it works. AI existed for years, but it was useless. Nothing useful, nothing useful, nothing useful. And then maybe last year or something, now it's really useful. There seemed to be some kind of jump in its ability, almost like a shock wave. We're trying to develop an understanding of how AI operates in terms of these shockwave jumps. Revealing how AI works will help society understand what it can and can't do and therefore remove some of this dark fear of being taken over and all this kind of thing. And If you don't understand how AI works, how can you govern it? To get effective governance, you need to understand how AI works because otherwise you don't know what you're going to regulate.
THE CREATIVE PROCESS
We talked about your curiosity growing up, but could you reflect on your upbringing—your parents, those teachers who have been important to you along the way? And as you think about the future and education, what would you like young people to know, preserve and remember?
JOHNSON
I was my first generation going to college. I've had some very inspirational teachers and some very non inspirational teachers, like everybody, but there was one that stands out by miles—a teacher I had in elementary school when I was eight. He was from Jamaica and just arrived in England. He was a jazz musician. He would start off the day playing jazz. He taught math as well, but he brought in a technique, which I still use to this day, to multiply numbers together. I've never seen it taught anywhere, it was done in some kind of slanted table. We'd done multiplication the previous year and I couldn't understand it, but he drew these tables and I thought this was just remarkable. That set me off and it showed me there was this interesting thing called jazz, and you don't have to do much to actually change a lot. Passing on some idea that might be unusual, as long as you can back it up, can have a really positive benefit going forward for future generations. So that's what I try to do. And I'm so happy for young people because I think the era of having to memorize things is going to go by the wayside. I think we're going to turn into a very creative period where education becomes creative, rather than jumping through the hoops and remembering things. I'm really excited for anybody who's entering there.
