Life with robots once seemed possible only in science fiction.
But today, scientific advances in artificial intelligence and robotics have ensured that robots are a part of our everyday lives.
On May 13, the Day of AI, we brought together a panel of experts to talk about the future of human and animal interaction with artificial intelligence and robots.
This is the first installment of our series Remarkable Science, featuring conversations with scientists about their discoveries, recorded in front of an audience at WBUR’s CitySpace venue in Boston.
Daniela Rus, professor of electrical engineering and computer science and director of the computer science and artificial intelligence laboratory at the Massachusetts Institute of Technology.
Merritt Moore, quantum physicist and professional ballerina.
Justin Werfel, senior research fellow and head of the Designing Emergence Laboratory, Harvard University School of Engineering and Applied Sciences.
How do you define AI and robotics?
DANIELA RUS: “It's important to know that there are three interconnected fields that get kind of jumbled and confused. So we have robotics, which puts computation in motion. We have AI, which gives machines the ability to make decisions. And then we have machine learning — which cuts across robotics and AI and many other fields.
"And as of now, machine learning is about taking data, analyzing the data, so that the machine can say what has happened in the future, what is likely to happen now, and what should you do in the future?
JUSTIN WERFEL: “My definition of a robot, I guess, is a machine that has autonomy. So nobody is driving it. It's reprogrammable. So it's not only limited, doing one thing, it can sense the world and it can act on the world.
" … AI is the thing, it is the brain of the robot, it's the thing that tells it, that goes between the sensing and the motion that says what it should do in response to what it senses.”
On the public misconceptions about the future of AI and robotics
DANIELA RUS: “Let me say that when I usually tell people what I do, I get one of two reactions. So some people start making jokes about Skynet and ask me when their jobs will go away. And other people say, ‘Great, when is my car going to be self-driving?’ So I actually belong to the second group.
"I'm very optimistic about technology. But in fact, the truth is that AI is not at the point of solving all of our problems. AI is not at the point of taking down the world. But AI has some very powerful capabilities that really augments what people can do. Both cognitively and physically.”
On possible guardrails that might be created to advise AI advancements
JUSTIN WERFEL: “I don't know what the right guardrails are, but I think the first step in finding them is for people to be asking the right questions. … People worry about whether robots are going to kill us. We could talk about why that's something to worry about, but I don't think that's something to worry about.
"Whereas, you know, when I have my self-driving car, people have a tendency to put too much trust in something that's not yet ready for it. … Understanding where the field is and where we should be thinking about putting guardrails, is going to be the first step in putting them in.”
MERRITT MOORE: “There is that fear that robots are going to replace humans. And my thoughts about it right now are like looking back in time when we had the painters and then the camera came out. And, I think every painter was like, ‘Oh my God, my job's over.'
"No one's ever going to want a painting ever again.’ And what turned out to be the case is that actually the camera became a tool for a different kind of human expression. So I think that I'm working with a robot, but it's not replacing humans. It's just a different tool for my expression.”
How do security and privacy play into your research?
DANIELA RUS: “We really live in a world where anybody can learn everything about us. So asking the question how we maintain privacy is super important. And here we can offer some technological advancements that explicitly target privacy. We are developing, in particular, two lines. One is in differential privacy. And so differential privacy refers to a body of algorithms that ensure that people can provide their data for the greater good, but then know individual data is localizable.
“Another direction is homomorphic encryption ... in homomorphic encryption we can do computations on encrypted data. And so then you can hold onto your data. You can, again, give it for computation for the greater good, but without revealing anything private about yourself. And so some of these technological solutions are getting us part of the way to privacy, but we need much more.”
For high income people, the robot is an opportunity. But for low income people, it could replace their jobs. How do we reach a balance?
DANIELA RUS: “Surprisingly, I would say that it's easier to send a robot to Mars than it is to get that robot to clear your dining table after dinner. So it turns out that there are a lot of tasks, primarily around manipulation, that our machines are really hard pressed to do well on. So I don't think we're anywhere near having all those lower income jobs go away.
“But having said that, I also think that it's important to anticipate what might happen in three years, in five years and put in place programs, educational programs, skilling and reskilling programs that enable people to get access to different jobs.”
On AI’s potential to help humans flourish creatively
MERRITT MOORE: "I think it would be beneficial to all of society if the doctors were able to sleep a little bit more and the robot was able to fill out that stuff. Right? And on the artistic side, I think we could come up with more creative ways to really become an expert in a specific field like ballet.
“I mean, many dancers are trained at age 6 and don't go to high school and [practice] eight hours a day for their entire lives. I think people would love to express movement in a way that kind of fuses the two together. But to be an expert in both is very difficult. …Then with robotics, it's just an interesting tool … I really find it as a tool to enhance human creativity.”
This interview has been edited for clarity. Highlights transcribed, written and compiled by Steven Davy.
Coming soon: A new series from On Point
Starting May 27, On Point will launch a four-part series exploring how artificial intelligence and machine learning may revolutionize the health care industry.
We'll investigate the technology already available or in development for clinical settings, criticize the ethical dilemmas the technology presents and understand the regulations in progress to advise AI advancements.
This series will also introduce listeners to the people involved in AI in health care; scientists developing tools, clinicians and doctors using the tools, and patients experiencing changing technology as part of their care.
This event is supported in part by Vertex, The Science of Possibility.
This segment aired on May 24, 2022.