In a recent essay in The Atlantic, Jonathan Merritt laments that theologians and Christian leaders, including Pope Francis, have not addressed what he claims will be the greatest challenge that Christianity has ever faced: Artificial Intelligence, or “AI.”
In his view, intelligent machines threaten to overturn many Christian beliefs, a trial that theologians seem blind to because “they’re stuck rehashing old questions instead of focusing on the coming ones.”
Such a criticism would be devastating if true, but is it?
A fuller reading of Pope Francis’s work suggests that he is actually engaging the issues with AI that most directly affect the contemporary Church and society. Before I get to that, though, it’s necessary to give Merritt’s argument his due. Most theologians are indeed not addressing the specific aspects of AI that he considers essential, but this is a wise choice on their part.
First, it’s important to note that “rehashing old questions,” or what Catholics like to call the development of tradition, provides many insights into these questions. For example, Merritt claims that “Christians have mostly understood the soul to be a uniquely human element, an internal and eternal component that animates our spiritual sides.”
This is not an accurate characterization.
Drawing upon the heritage of Greek philosophy, most theologians have understood the soul to be what makes a specific living thing what it is. It is the principle of growth and development in all living things, movements and sensation in animals, and rationality in humans.
Therefore, animals have souls, plants have souls, and an AI that could think and manipulate the world around it would have to have something like a soul.
Merritt qualifies himself in the next sentence to refer to the image of God that each person possesses in her soul. Yet again, major figures in the tradition such as Thomas Aquinas do not see the image of God restricted to humans.
For him (some other theologians have very different interpretations), we imagine God primarily in our potential for reason and free will, so any being with reason and free will would possess that image, including angels, for Aquinas, rational aliens, for Francis, even true AI, if it existed.
Of course, this reason is not mere instrumental reason, but one that understands purposes, meaning, and the moral law.
Still, based on Merritt’s argument one might ask, how can such spiritual faculties arise out of silicon circuits (or nanotubes, or any other material)? While a problem, it is no more difficult, nor much different, than the question of how the spiritual arises from lowly flesh, a question that thinkers have wrestled with throughout the Western tradition.
Theologians struggle with this problem in ordinary human development – how and when new life gains a soul is a central theological question, for obvious practical reasons. The predominant answer in the Catholic tradition is that, in the process of procreation in which human parents cooperate, God creates an individual spiritual soul for each human body. Something like this framework could be used to think about AI.
It is true that some issues are more difficult, like how AI could be redeemed.
Christianity argues for God’s special care for humanity, with the second person of the Trinity assuming a human nature in the Incarnation. This doctrine raises questions about Christ’s relation to any possible AI, but ones not fundamentally different to questions of how Christ redeems all of nonhuman creation, questions that have become ever more pressing given environmental devastation.
Given these resources, why haven’t more theologians directly addressed AI?
First, I would guess that most theologians are less optimistic than the ones Merritt quotes about the actual possibility of true AI. Beyond the sixty years of unfulfilled promises that AI is just around the corner, AI theorists have not addressed philosophical concerns as to whether their programs can have consciousness and grasp meaning.
In his Chinese Room argument, John Searle pointed out that while computer programs manipulate symbols (syntax), allowing them to imitate behavior, they cannot really grasp the meaning (semantics) of the things they manipulate, which would be necessary for consciousness.
A second source of skepticism for engaging AI is that, along with many contemporary non-Christian thinkers, theologians recognize making an AI is an extremely bad idea. If a machine has the free choice necessary for true AI, then it has the possibility of sin, leading to large downside risks, such as human extinction.
This concern about risk raises the final problem with Merritt’s analysis – if one reads Francis carefully, one finds that he addresses the problems of today’s limited AI that are harming people right now rather than future speculative possibilities.
Laudato Si’, Francis’s recent encyclical, is just as much about technology in human ecology as it is about the natural environment.
He addresses contemporary mental pollution and isolation, reflecting concerns in other papal addresses over people only receiving information that confirms their opinions, problems that arise in part due to AI algorithms reflecting our opinions back to us in search results and news feeds, a solipsism whose political effects were chillingly documented in Adam Curtis’ documentary HyperNormalisation.
In a second and even more important example, he laments “a kind of technological progress in which the costs of production are reduced by laying off workers and replacing them with machines.” These are not only issues of automation impacting blue collar jobs, but now, even many white collar jobs are disappearing due to the applications of AI.
Pope Francis demonstrates that dealing with Merritt’s speculative problems may distract us from more pressing challenges, such as knowledge workers in their late 40s whose positions become redundant due to AI and who thus won’t be able to make their mortgages while they retrain.
Problems like that may not be as hot a topic for a TED talk as speculating on the prayer life of AI, but these are the challenges of technology that a Church whose members will be judged by their care for the least in society should be addressing.
Paul Scherz is an assistant professor of moral theology/ethics at The Catholic University of America. He examines how the daily use of biomedical technologies shapes the way researchers, doctors, and patients see and manipulate the world and their bodies. Scherz has a Ph.D. in Genetics from Harvard University and a Ph.D. in moral theology from the University of Notre Dame.