How AI Will Change Us
Dynamism of a Human Body, 1913 Umberto Boccioni (1882-1916 Milano, Castello Sforzesco)
Artificial intelligence, in various forms, is increasingly and impressively a part of our daily lives. It guides the advertisements that we see, the routes that we take, the distribution of power in our cities, the allocation of medical care in our hospitals—and, of course, it responds to the questions that we ask our smartphones. Soon it will drive our cars. In some ways, to ask how it will change us is to come late to the game; it has already significantly changed our ways of thought and our way of life, often quite invisibly. To consider its present and future impact, Part I of this essay will survey a brief history of AI, reflect on its current status, and explore four AI-related questions: The nature of the mind; the relationship between humans and machines; the impact of machines on human welfare; and the impact of AI on society. Then, in Part II, we will consider social AI—that is, AI that simulates human sociality—in order to highlight the possibilities and perils of our near-term future.
AI and Our Present Situation
First, a quick history of artificial intelligence, or AI. ENIAC, the first general-purpose computer, was created at the University of Pennsylvania in 1946. Four years later, in the midst of this early work on computers, Alan Turing, a British computer scientist who played an important role in breaking the Nazi communications code during World War II, wrote a paper in which he discussed how to build intelligent machines and proposed a test for determining whether a computer has artificial intelligence, which amounted to whether it could pass itself off as a human being in conversation. At the Dartmouth Workshop on AI in 1956, prominent researchers from MIT, Carnegie Mellon University, Bell Labs, and elsewhere gathered to explore whether human intelligence could be simulated by a machine. Since that seminal meeting, there have been many ups and downs, but overall progress has been made in various AI-related fields. Among the most prominent are natural language understanding, autonomous driving, customer recommendation services, and computer vision. Machine learning is an important aspect of AI, and the fairly recent advent of deep learning has led to swift advances in some difficult applications.
Researchers now distinguish between “narrow” and “general” AI. “Narrow” (also called “Soft” or “Weak”) AI refers to the ability to perform simple or specific tasks. An example is the ability to play a specific game such as Go, or the natural language understanding in Siri or Alexa. “General” (also called “Hard” or “Strong”)* AI refers to human-like intelligence, particularly the ability to deal with ambiguity, exhibit intuitive understanding, and handle and switch between a variety of different contexts and situations without becoming confused. Artificial General Intelligence systems don’t exist as yet, but creating them has been the goal of many AI researchers since the field was founded.
What, then, can AI currently do? AI-based computer intelligence has already displaced the world champions in the games of Chess, Go, and Jeopardy, specific tasks that require the ability to rapidly sift through and analyze large amounts of information. Google’s search engine can perform a similar task with a broader set of information for news, commerce, research, and other purposes. AI-aided computer vision algorithms can recognize faces and perform scene understanding for autonomous driving. And the company Boston Dynamics recently released a video of their robots performing funky dance moves a lot of us would envy. Each of these is a good example of “soft” or “weak” AI. They are impressive, but narrow—they don’t exhibit the full range of human capabilities or characteristics.
Initial Questions and Themes
Now that we have a common understanding of the basics of AI, let us pose four related questions, outline their implications, and make provisional comments on each of them as a basis for further discussion. These questions are: What is the nature of the mind? What is the relationship between humans and machines? What is the impact of machines on human work? What is the impact of AI for better or worse?
What is the nature of the mind?
This question is fundamental because it is tied up with one’s view of whether “an AI,” i.e., a computer-based human-like being or “android,” can eventually achieve something like human status. A related question from the standpoint of Christian theology is, what constitutes the image of God, or imago Dei, taught in the Bible starting in the Book of Genesis? Theologians and other thinkers have given various answers, and the possession of intelligence, or a mind and its accompanying power of rational thought, is one of them. For the moment, let us proceed as if that were the correct answer, and see what it may entail.
The materialist who denies both God’s existence and the Christian idea that humanity has a unique status within the created world might argue as follows: “Christians claim that humankind is special in that it bears God’s image, and this image is evidenced by the presence of mind. But computer scientists have created, or will someday be able to create, a machine that appears to have a mind. Therefore, humankind’s supposedly unique status is disproven.” On the other hand, a Christian might argue as follows: “God gave the human being a mind as part of creating him in His image. Humanity in turn partakes of God’s creative ability and is thereby able to impart to a machine the appearance, or perhaps even the reality, of mind. Humankind’s uniqueness in possessing God’s image is therefore upheld by his ability to reflect that image in what he himself creates.” Who has the better of this argument?
In order to address this, let us think for a moment about the idea of rationality. From the standpoint of the materialist, i.e., a person who believes that the only thing that exists is matter, rationality has arisen from a long evolutionary process of random trial-and-error guided not by the objective of achieving truth, which cannot possibly inhere in matter, but by the objective of gaining competitive advantage in the struggle for survival. The materialist trusts in the ability of the undirected motion of molecules in the brain to accurately conclude that that’s all that’s happening in the brain, a self-defeating and contradictory conclusion. In other words, if materialism were true, we would not be able to confidently reason that it is. The result of this argument is that, to be dependable, rationality cannot rise by itself from nothingness, matter, or randomness; rather, it needs to be underwritten by or grounded in something or someone outside of the material universe.
Returning, then, to the argument between the materialist and the Christian. Is the materialist right to say that there is nothing special about the human being because the mind can be imitated? A possible Christian response is that the mind, with its ability to reason truly about the material and immaterial worlds, has a non-material origin; therefore, even if a mind can be imitated or artificially created by humans (in itself a debatable proposition), such a mind has an ultimately transcendent, and therefore special, origin.
To round this point out, note that many theologians consider aspects other than mind as part of the imago Dei, such as the ability to form relationships and have a subjective experience of self-possession and self-donation. AI researchers have worked on giving apparent emotions to robots, but we seem no closer to giving real emotions to robots. There are, in the end, more aspects to the question of creating an AI than simply mind, however mind may be construed. Finally, some Christians, such as the British physicist and neuroscientist Donald MacKay, who attended the 1956 Dartmouth Workshop on AI, have proposed that if humans were able to create a machine that seemed fully human, then the machine’s responsibility would be to worship God as its ultimate, rather than intermediate, creator.
What is the relationship between humans and machines?
Humans have built machines for millennia in order to make work easier and to enable us to do things that we could not otherwise do. For the most part, this has been viewed as a positive application of our creativity and has not raised ethical eyebrows. No one complains that hearing aids or hip implants raise the question of what it means to be human. But what about the prospect of a chip implant that joins one’s brain to the Internet? We think with our digested memories, and so instant access to whatever we may seek will not amplify our knowledge but only the availability of undigested information. A chemist accomplishes more with a periodic table than does a computer scientist—and so, I expect, we shall still have to develop a habitual familiarity and deep reflection upon the information to which we have access if it is to really alter the way we think about things. But what if, somehow, we were able effortlessly to achieve some kind of admittedly tawdry earthly omniscience? What implications would that have for the perennially lauded virtues of study and hard work to achieve knowledge?
The extreme version of the drive for augmentation is transhumanism, a movement among some AI researchers and technologists that seeks to use AI and related technologies to dramatically modify our own intellectual and physiological capabilities. One of its ideas is that human consciousness or intellect is separable from the body and can be “downloaded”—i.e., electronically copied—into a more lasting, “non-carbon-based” body, with the intent of granting immortality. Thus, transhumanism grates on Christian sensibilities in at least two fundamental ways: it seeks to be like God in a way reminiscent of the biblical story of the Tower of Babel, and it denigrates the human body in a way reminiscent of Gnosticism. Moreover, it remains unclear how my own experience could be contiguous with that of the new electronic intellect—i.e. how this would be my immortality rather than the immortality of (at best) a copy of myself. Whatever the case, the question of to what extent we should use technology to aid and augment humans is a difficult one and not answered simply by dismissing the transhumanists.
What is the impact of machines on human work?
On the one hand, machines ease and enable work, and many people enjoy working as a way to exercise their physical and intellectual powers and to create lasting artifacts that enrich human life. On the other hand, as machines become more capable through AI, they are able to supplant a greater number of workers at higher skill levels. Though historically job creation has more than kept pace with job loss through technology, today many express the concern that there will be unusually rapid replacement of entire classes of workers without the possibility of retraining for new positions, many of which require greater technical sophistication than in the past. An example is the technology of autonomous driving. There are as many as 5 million professional or part-time drivers in the U.S., including 3.5 million truck drivers. Based on the slower than expected progress of the technology, especially with respect to guaranteeing an acceptable level of safety, these jobs will not disappear overnight, so there is more time for adjustment than the most alarmist commentaries suggest. However, this example and others highlight the need for careful thought and action regarding education, job re-training, and economic relief policies. Finally, the ability to work and, from a Christian perspective, to be good stewards of creation is tied up with human dignity and significance, so technology that pushes us in the direction of total leisure should be resisted.
What is the impact of AI for better or worse?
First, it’s clear that AI, like any powerful tool, has the potential for great good as well as great evil. In addition to the goods cited above, we might think of using AI to identify and interdict terrorists via face or activity detection. But the same AI could be (and is) used by totalitarian governments to control and spy on their own people. And there are many intermediate examples, such as the advantage of being able to make convenient purchases on Amazon being coupled with the downside of your personal preferences being captured in order to target you for future purchases or sell your information to other companies. Second, there is a significant and encouraging amount of idealism and activity among students and fellow researchers regarding the use of AI for social good, and concern about its potential abuse in policing, social control, and other areas. In all, the nature and meaning of our own personhood becomes a necessary object of examination wherever some technology is deeply and transformatively entwined with our lives—but this is particularly so when that technology imitates or purports to replicate what we are.
*Note: Some scholars reserve the adjective “Strong” for AI systems that would not only have similar competence to human intellects, but would exercise that competence by the same means, i.e., including phenomenal consciousness and subjectivity. A “Strong” system would thus not only match a human’s behavioral agency but also would be a replication of how the human did it.
John M. Dolan received his doctorate in mechanical engineering from Carnegie Mellon University, where he has been a faculty member in the Robotics Institute since the 1990s.
Jordan Wales received his doctorate in theology from the University of Notre Dame, and has been on the faculty of Hillsdale College since 2014.