Our Robots, Ourselves: Robotics and the Myths of Autonomy
Our Robots, Ourselves: Robotics and the Myths of Autonomy by David A. Mindell. Penguin Group, 2015. 260 pages.
From drones to Mars rovers – an exploration of the most innovative use of robots today and a provocative argument for the crucial role of humans in our increasingly technological future.
In Our Robots, Ourselves, David Mindell offers a fascinating behind-the-scenes look at the cutting edge of robotics today, debunking commonly held myths and exploring the rapidly changing relationships between humans and machines. Drawing on firsthand experience, extensive interviews, and the latest research from MIT and elsewhere, Mindell takes us to extreme environments-high atmosphere, deep ocean, and outer space – to reveal where the most advanced robotics already exist. In these environments, scientists use robots to discover new information about ancient civilizations, to map some of the world’s largest geological features, and even to “commute” to Mars to conduct daily experiments. But these tools of air, sea, and space also forecast the dangers, ethical quandaries, and unintended consequences of a future in which robotics and automation suffuse our everyday lives.
Mindell argues that the stark lines we’ve drawn between human and not human, manual and automated, aren’t helpful for understanding our relationship with robotics. Brilliantly researched and accessibly written, Our Robots, Ourselves clarifies misconceptions about the autonomous robot, offering instead a hopeful message about what he calls “rich human presence” at the center of the technological landscape we are now.
Overview
Our Robots, Ourselves: Robotics and the Myths of Autonomy, by David Mindell, frames autonomy from a different perspective. Autonomy isn’t necessarily the incorporation of advanced electronics and artificial intelligence. Mindell writes, autonomy is “human action removed by time.” A human performs the task by physical presence or by program in the laboratory. Despite the distance or number of intermediary steps (between the programmer, operator, or physical action), the human is still present; albeit remotely. Once one thinks of autonomy this way, many of the dilemmas it presents fall away. Automation of a task does not alter its nature fundamentally, though it may alter the way it’s perceived.
Mindell uses several examples of autonomy, starting with a robot named “Jason” mapping the seafloor. This example is straightforward and non-controversial – there is a device doing long endurance work using sophisticated sensors in a dangerous environment. More controversial are military drones. Readers with military backgrounds may understand the basic employment of unmanned aircraft. Mindell’s narrative uses the evolution of unmanned aircraft from guided missiles to Predators and Reapers as a step towards his larger point.
This book also examines a related issue in manned aircraft. Modern airplanes can operate autonomously for most of their flights, but humans are still physically present, and for now, in charge. It is in this example the complexity of man-machine interaction is clearest. While many aircraft can fly thousands of miles automatically, humans still make inputs periodically. When they do that, they often make incorrect assessments of what their flight director is still doing for them. This has repeatedly caused mishaps and near-mishaps. In many ways, the increasing complexity of modern cockpit automation has made that interface more complex, even if the goal is to make aircraft operation simpler.
The aircraft example gives relevant insight to a way forward. At certain airports, airplanes can land, even in zero visibility. Pilots, however, can select command guidance that allows them to see where the computer-optimized control inputs are. So, if they want to, they can manipulate the controls to “put the thing on the thing” or they can add their human judgment to deviate from it.
This epitomizes the thrust of the book. The challenge is how to make humans and machines work together most effectively. The issue is not if things will be automated, but how. Machines will not only change how tasks are done, but they will change the tasks themselves. Marine archaeology, flying, driving, space exploration, and other fields Mindell discusses will not look the same as we thought they would twenty years ago, ten, or even today.
Takeaway
Mindell’s book doesn’t give an academic framework on the science or ethics of automation. But it does expose the reader to several examples of automation and different ways of thinking about them. This is not a detailed work on how autonomy affects command and control or rules of engagement. What it does is challenge the way one thinks about autonomy and is a useful introduction and steppingstone to more specific discussions. An expert on autonomy may not need, or at least not think they need, this book. But the novice does. The expert probably needs it more than they think.
