The Coming Psycho-Social Challenge of Advanced Robotics
The Coming Psycho-Social Challenge of Advanced Robotics.
Two weeks ago, Bill Bowen wrote a very thoughtful essay reflecting on why the U.S. and other advanced industrial societies are experiencing so much widespread social anxiety, divisiveness, and instability today.
Although broad social trends have many causes, Bill’s essay highlights an important factor that is often overlooked. We are living in an era of increasing dependence on complex socio-technical systems (STS). As our dependence on STS’s grows, the psycho-social effects of that dependence can foster a wide range of problems in social relations that need to be addressed.
In this essay I want to expand on Bill’s discussion. If he is correct (and I firmly believe he is), we better develop comprehensive public policies to address these new psycho-social problems quickly, because I believe we are at the very early stages of an even more radical transformation.
Although many aspects of technology have advanced aggressively in the last several decades, perhaps the most consequential has been the so-called Digital Revolution. This revolution has three core aspects.
The first is the capacity to create complex digital representations of the physical world along with digital representations of our own social and emotional interactions with the physical world. This combination captures in digital formats the human processes of how we interpret the world and how we create meaning in our lives.
The second aspect of the Digital Revolution is the explosive growth in our capacity to gather, store, and process enormous amounts digital data. The third is the equally explosive growth in the scale and the scope of instantaneous communications that has become routine among almost all data processing and storage systems.
The net result of these three aspects of the Digital Revolution transforms billions of interconnected streams of data gathering and the construction of digital representations into real-time continuous flows of communication in visual, audible, and even tactile modes. These continuous flows increase our complex dependence on the socio-technical systems of daily life in ways that create many benefits.
Yet they also sew new forms of psychological confusion and alienation. They disrupt our capacity to create meaning by making it difficult to distinguish between the physical world that we know through our biological senses and the virtual world that is created by our digital representations. Much of the social anxiety, divisiveness, and instability that Bill Bowen discussed can be connected, at least partially, to these aspects of the Digital Revolution.
As we approach the end of the first quarter of the 21st century, I think many of the remedies that Bill Bowen suggests have great merit. We need to guard against relying too much on virtual representations of the world to inform meaning. We can achieve this goal by engaging with other people in community organizations, by building our social connections with diverse groups of people, and by seeking out opportunities to collaborate with others in mutual problem solving.
These actions can reinforce our capacity to think critically, and creatively, about conditions in the physical world, and thereby sharpen our critical interpretations of the virtual worlds that we encounter through digital representations. Yet those remedies assume there is a wall between the physical world, i.e. the “real world,” and the various virtual worlds that we can experience through digital communication.
As of today, digital representations of the world are just that. They are representations. They are partial renditions of the physical world that convey some type of meaning that is likewise partial and select. They exist separately from the complex dynamics of the real physical world. They do not interact with the real world directly.
Virtual worlds currently connect to the physical world only if they convey a meaning that inspires people to become active in the real world. An impressionistic painting of a well-tended city park, for example, may convey the park’s beauty and tranquility very well. Yet it will exclude many components. And the painting itself has no effect on the park’s beauty or tranquility unless it inspires someone to intervene in the park’s actual physical conditions.
Yet the walls between the physical world and the virtual world are becoming porous. Within our lifetimes, likely much faster than we may want, those walls will tumble entirely. The rapid development of robotics technologies will begin connecting virtual worlds to the physical world without the need for people to function as actors in between.
Robotics is all about the integration of different types of technologies to create virtual representations of desirable conditions in the physical world and then actively intervene in the physical world to restructure it so that the physical world conforms to the desired virtual delineation.
Although the field of robotics seeks to integrate many technologies, the integration of five core technologies defines most of the field. They are mechatronics, sensors, servomechanisms, virtual reality, and artificial intelligence. Each core technology has developed on its own trajectory over the last several decades, but the Digital Revolution has transformed them all and made them capable of more seamless integration.
Mechatronics is a term that describes the integration of mechanical engineering, electrical engineering, software control, and numerous other specialties, to create durable, functional devices that can manipulate specific items in the physical world.
Sensors are devices that produce output signals that contain information about any type of specific condition in the physical world. They can use any part of the electromagnetic spectrum to convey aspects of heat, light, motion, moisture, pressure, geometry, relative position, or any other aspect of a specific environment’s condition.
Servomechanisms are electronic devices that receive inputs about the physical environment from sensors and send outputs to mechatronic and/or other types of devices that move or act within that environment for the purpose of manipulating objects to achieve desired results or achieving specific environmental conditions within that space. Servomechanisms operate dynamically to achieve desired outcomes by continuously monitoring and interpreting feedback and continuously translating feedback into instructions to cause change within physical space to achieve some desired outcome.
Virtual realities, as discussed previously, are partial digital representations of complex physical spaces. They are engineered to portray conditions that represent selected desired outcomes that convey desired meanings. In the world of robotics, virtual realities define the goals that are used by servomechanisms to govern the continuous process of generating iterative instructions to intervene in the physical world and to assess sensory feedback.
Artificial intelligence is the least developed of the five core technologies of robotics, as of now. The role of artificial intelligence in robotics is to define optimal desired outcomes for any specific environmental setting. In addition, artificial intelligence seeks to provide continual oversight of the interaction of the other four technologies to optimize the effectiveness and efficiency of the strategies used to achieve desired outcomes.
In simplified terms, artificial intelligence creates optimum virtual realities and produces optimal algorithms that servomechanisms can use to strategically use mechatronics and other devices, in tandem with sensors, to manipulate the physical world so that aspects of the physical world are changed to conform with desired, optimum virtual realities. In many settings the field of robotics has evolved to create successful systems integrations that routinely produce excellent outcomes, despite the limited state of artificial intelligence today.
Millions of robotic systems around the world already integrate core technologies to build high-quality products ranging from simple plastic toys to complex automobiles, to the most sophisticated computer chips. State-of-the-art forms of artificial intelligence (expert systems, machine intelligence, machine learning, etc.) provide greater and greater self-management and improvement to these industrial production systems.
Other examples of sophisticated robotics systems go far beyond controlled production settings. Self-driving vehicles, both cars and trucks, and automated earth moving equipment are examples. Robotic surgical systems are other advanced examples. Robotic surgery is now common in urology, gynecology, cardiothoracic surgery, neurosurgery, and orthopedics.
Even the most advanced robotics systems, however, operate within defined boundaries by using rudimentary artificial intelligence capacities within limited foundational models of their defined tasks. Broader advances within artificial intelligence, such as the GPT general language models, represent important breakthroughs but still fall short of the field’s long-term dream of a general-purpose artificial intelligence capacity that can operate in almost any problem-solving context without the need for interactions with human intelligence. This capacity, which is known as polymathic artificial general intelligence, is evolving much faster than previously predicted.
Earlier in my career (1998-2002) I learned a lot about robotics when I was helping to put together an economic development strategy to make Pittsburgh the center of commercial R&D for robotics. One of the biggest skeptics I encountered then was Herbert Simon, who was one of the founders of the field of computer science and is often cited as the principal pioneer of artificial intelligence.
Simon advised me that it could take a century for researchers to develop a truly general-purpose thinking machine that could outperform the human mind, if that was even possible. And it would take even more time to develop machines that could function autonomously in the world with that level of cognitive power.
In contrast, the New York Times columnist Tom Friedman recently interviewed Craig Mundie, who served as Microsoft’s most senior research and strategy officer. Friedman quotes Mundie’s assessment, “It is quite conceivable that we will achieve polymathic artificial general intelligence in the next three to five years, so it is also likely that our next president, and certainly the one after, will have to cope with the fundamental societal changes that will result.”[i]
What will be the psycho-social consequences for people when the complexity of the socio-technical systems that are already integral to our daily lives increases dramatically by the introduction of new robotic systems that are powered by artificial general intelligence capacity?
How will we cope with new robotic systems that may have the capacity to operate autonomously in the real, physical world to make the physical world conform to optimal virtual realities defined by the robotic systems?
How will these new systems reform human social relations, and how will humans maintain the capacity to enforce some form of democratic human governance over this transformed form of society?
No one knows the answers to these questions. But we are all likely to live through the process of finding out. In future posts I’ll discuss some of the potential answers that are being explored.
Bob Gleeson
[i] Thomas A. Friedman, “A Harris Presidency is the Only Way to Stay Ahead of A.I.” NYT, October 29, 2024.