recent

5 ‘extraordinary turnarounds’ for global well-being

Close proximity is the key to knowledge spillovers among startups

6 ways to nurture IT talent

Credit: LeArchitecto / angkhan / iStock

Ideas Made to Matter

Robots

Robots aren’t magic, but they could give humans ‘superpowers’

By

There is a robotics revolution underway, according to Daniela Rus, a prominent roboticist and director of the MIT Computer Science and Artificial Intelligence Laboratory. As she points out in a new book, there are now 3.1 million robots working in factories — packing goods, assembling products, monitoring air quality, and more.

Written with science writer Gregory Mone, “The Heart and the Chip: Our Bright Future With Robots,” lays out a vision for how humans (“hearts”) and robots (“chips”) can work together for powerful new applications and capabilities. “If this revolution is correctly and intelligently steered, smart machines have the potential to improve the quality of human life as dramatically as the plow,” she writes.

In this excerpt from the book’s introduction, Rus looks at the technologies underpinning robots, what humans still do better than machines, and how robots could give humans “superpowers.”

The excerpt has been edited for style and length.

+++

Let me clear up a common misconception. Despite what you see in the movies or read about artificial intelligence online, robots have not transformed into magical, all-powerful, independent entities. They’re also not evil forces to fear, as science fiction so often suggests, implying that intelligent machines will inevitably revolt against their human maker.

Robots are tools. They aren’t inherently good or bad. Neither is a hammer. Think of this new generation of extraordinary machines as very advanced hammers; it is what we choose to do with these tools that defines their impact and value. And we can choose to do incredible things. We can work alongside robots to engineer better medicines, make transportation safer and more efficient, assign them to tasks that are too dangerous or difficult for humans alone, translate conversations instantly into other languages, and even give ourselves superpowers.

Yes, superpowers. I am a very proud robot mother. What the field as a whole has accomplished in the past decade is stunning, but what we are poised to do in the next 20 years is even more thrilling. Truly, we’re just getting started. The recent breakthroughs in AI are only the beginning.

Yet robots are still primitive compared to humans. Working on these technologies has only amplified my appreciation for the mental and physical marvel that is the human being. The soaring voices of opera singers, the majestic, gravity-defying movements of ballet dancers, the incomparable beauty of a powerful poem, or the simple elegance of an equation that reduces a law of nature to a few connected variables — such feats will remain the work of people.

Indeed, researching AI and robotics has given me a deeper understanding of our specie’s incredible cognitive, imaginative, and creative powers — not to mention our physical strengths. We do so, so many things better than robots.

Yet intelligent machines surpass us in some areas. Thanks to their powerful chips, machines are capable of churning through and processing massive volumes of data at speeds people could never approach. Robots can be built to carry out specific tasks repeatedly with phenomenal precision.

But even the most advanced machine intelligences lack wisdom, knowledge, and understanding. They don’t deal as well with uncertainty or unforeseen changes. They can mimic art, but they aren’t nearly as creative as artists. A machine might be able to produce something that loosely resembles a Picasso, but there will be no Picassobot, no robot that channels the current ideas at large in society to generate a new and groundbreaking style of expression like cubism.

An AI-powered writing program can now generate text that seems to be sensible or even indistinguishable from human-authored writing, but it will not produce great or lasting work that offers insight into the mystery of the human condition. There will be no AI Shakespeare or Tolstoy. Robots have amazing capabilities, but they fall short in many areas, and they lack the intrinsic, innate drive to care. In short, they lack heart.

Too often we manufacture tension between people and robots, or the heart and the chip, when we should be thinking more about how to get them to work together to optimize the strength of each. Humans steer the design of this technology and its applications. What happens with robots and AI is up to us. And when we focus on combining the strengths of robots with those of humans, or pairing the heart and the chip, the results can be astounding.

Some years ago, researchers conducted an experiment in which trained humans and bespoke machines were asked to study images of lymph node cells and determine which ones were cancerous and which ones were not. The machines had an error rate of 7.5%. The humans had an error rate of 3.5%. So the heart still beats the chip. But here’s the more exciting point: When the humans enlisted the help of the machines and the two worked together, the error rate dropped to 0.5%. That’s an 80% improvement in diagnostic accuracy.

Imagine a future in which every health practitioner — even those working in small practices, in rural settings — has access to these solutions. A hardworking doctor cannot possibly have the time to review every new study and clinical trial. But working in tandem with these systems, the doctor will be able to offer patients the most cutting­-edge diagnosis and treatment options. And robot- and AI-enhanced advances like this one could be put to work in almost every professional field.

My aim is to help you understand the essential distinctions between humans and our intelligent machines and to help you see how and why working closely with robots, and channeling the strengths of the heart and the chip in the service of people, will allow us to build a better — and more exciting — future for everyone.

The technologies I focus on draw from three interconnected fields. The first, robotics, puts computation in motion by giving a computing system a physical, mobile body. Imagine your smartphone with wheels, wings, or even a hand. Artificial intelligence, or AI, empowers machines with the ability to reason and make decisions in very specific, focused areas. Competing at chess, for example, is an example of what we call narrow AI, which is the dominant form in the world. General AI, the kind that guides all-powerful movie robots? That remains a distant possibility. A task-oriented AI that plays chess like a grandmaster, for example, couldn’t steer a car through an intersection or help a robot pick out a coffee cup on a countertop.

The third field, machine learning, cuts across robotics and AI to look at large stores of data, find patterns, and make predictions or formulate conclusions with some degree of confidence. This is what enables a program to scan millions of images of trees, then look out at the world and identify a tree it has never seen before.

In the radiology example above, a machine learning system identified patterns in the images that matched patterns it had identified in previous images of cancerous lymph nodes. That pattern didn’t mean the individual had cancer, but it nudged the medical professional to take a deeper look at certain scans and thereby reduced the margin of error.

Related Articles

Measuring the actual impact of robots on jobs
How smaller firms can integrate collaborative robots
How companies can pursue ‘positive-sum automation’

Machine learning and artificial intelligence are often confused, as the latter has become a business and marketing buzzword, but you can think of machine learning as pattern recognition systems that work in service of AI, to assist with higher-level decision-making and reasoning. Where machine learning and AI fall short is in understanding what cancer or a tree actually is.

At an early age I started imagining how we might be able to surpass our biological limitations to do more than the body allows. Later I wanted to scale buildings like Spider-Man, fly like Iron Man, shape-shift like Mystique, turn invisible like Sue Storm of the Fantastic Four, or be as strong as Superman. These superpowers seem to have always existed in storybooks, abilities we wish to have but cannot have.

But robots and computation could make these phenomenal capabilities more accessible to all of us. Robots move and accomplish tasks independently, saving us time. We could work with them to enhance our perception, reach, precision, strength, and ability to process and respond to huge datasets. Robots could also give us entirely new capabilities that would have seemed impossible in years past. What seems like magic will arise from the clever and creative use of mathematical models, algorithms, ingenious designs, new materials, and electromechanical components.

Robots really could give us superpowers.

Excerpted from “The Heart and the Chip: Our Bright Future With Robots.” © 2024 by Daniela Rus and Gregory Mone. Used with permission of the publisher, W.W. Norton & Co. All rights reserved.

For more info Sara Brown Senior News Editor and Writer