AI is no match for the marvels of humanity

Magnifica Humanitas, Pope Leo XIV’s encyclical on artificial intelligence, has accelerated the debate on the fundamental nature of the technology. Could AI have a conscience? How does it think? What does its rapid advance portend?
In the encyclical’s most provocative section, Leo argues that AI systems “do not undergo experiences,” “do not feel joy or pain,” and do not “have a moral conscience, since they do not judge good and evil, grasp the ultimate meaning of situations, or bear responsibility for consequences.” He continues: “They may imitate language, behavior and analytical skills, or even simulate empathy and understanding, but they do not understand what they produce.”
Whether you agree with the pope’s interpretation depends in part on your answers to foundational questions about what it means to be human and the nature of consciousness.
At the presentation of the encyclical in Vatican City, Anthropic co-founder Chris Olah claimed that his research team, which studies the internal structure of these models, has found “evidence of introspection” and “internal states that functionally mirror joy, satisfaction, fear, grief, and unease.” While Olah admitted that he does not know what that means, he posits that “it warrants ongoing discernment.”
Well, I would like to see some evidence. The burden of proof that AI can engage in introspection — can form a self-image, as humans can — and experience emotions is on the believer, not the skeptic.
What about cognition? Leo argues that AI tools “merely imitate certain functions of human intelligence,” and that they are “entirely tied to data processing.” Many technologists disagree. But the pope is right to highlight a distinction between AI’s data processing and human cognition. Generative AI tools excel at pattern recognition. The statistical models that power them use an inductive approach, relying on huge data sets and massive computing power to imbue AI systems with tacit knowledge.
This differs from human learning in important ways. We do not train our minds on enormous quantities of data with the goal of using these inputs to predict outputs. Instead, we theorize and hypothesize based on a small number of examples, often from our own experience. We are tribal, learning from our families and communities, often adopting the conclusions of those around us. We engage in trial and error. 

The pope focuses on how AI could undermine human flourishing.

Michael R. Strain

As fears about the potential effects of generative AI grow, Leo is admirably clear: Humanity’s creative intelligence, including AI and any other major technological advancement, “is a gift that can alleviate suffering and open up new possibilities, but it must remain ordered toward the common good, justice, the care of the vulnerable and creation.”
The pope focuses on how AI could undermine human flourishing, whether by ushering in autonomous warfare, exacerbating inequality, violating privacy, causing mass unemployment, or reducing human beings to cogs. We may even gradually lose “the very desire to form genuine human connections.”
To be sure, acknowledging the risks is prudent. But I wish that the pope had spent more time on AI’s enormous potential to improve human welfare. These tools could accelerate pharmaceutical innovation; enhance the quality and accessibility of healthcare and education, particularly in the developing world; create jobs; and raise living standards. Indeed, I expect that, on balance, AI will substantially increase human flourishing.
I also wish that Leo had more faith in markets and less in government. Instead of recognizing that AI will likely increase the working poor’s earnings, he seems more concerned about how policymakers can counter AI-driven inequality, and how private corporations will use these tools to grab market share. Leo calls for “measures to ensure equity,” including “industrial policies,” with apparent confidence that they will reduce the concentration of wealth and power, when they are more likely to have the opposite effect.
Ultimately, these disagreements are less important than the moral and intellectual foundations of Leo’s first encyclical. In a world captivated by technological advances, the pope emphasizes the primacy of each person’s inherent and inestimable dignity. And amid the frenzy around picking AI winners and the hype about AI possibly eclipsing humanity, Leo asks us to protect the common good and challenges us to embrace our weakness and frailty. As he puts it: “We must remember that humanity flourishes not despite limitations, but often through them.”
In Magnifica Humanitas, Leo seeks to shift attention away from the marvels and terrors of AI and toward the magnificence of humanity. With all eyes now on the technology, this message is needed and welcome. AI tools are impressive. But they are pedestrian compared with the grandeur of a human being.

• Michael R. Strain, Director of Economic Policy Studies at the American Enterprise Institute, is the author, most recently, of ‘The American Dream Is Not Dead (But Populism Could Kill It)’ (Templeton Press, 2020).
©Project Syndicate