Why Religious Leaders Warn about AI

Google Image

Is there a subject about which the media has been more obsessed than Artificial Intelligence (AI)? Daily, news and commentary about AI is in the newspapers, on TV news and social media. Most of us are struggling to understand it – including those of us who use it frequently.

Surprisingly, religious leaders are weighing in. Jay Kim, a Protestant pastor in Silicon Valley, has warned that Christians tend to regard digital technologies as neutral tools when they actually shape their users. Discussing AI, he emphasized its potential to be dehumanizing and to substitute artificial relationships for embodied human relationships.

Not to be Confused with a Person

Rabbanit (a title for a female rabbi) Sara Wolkenfeld, chief learning officer of Sefaria, an online library of Jewish texts, has emphasized a different danger: AI can feel like a person even though it isn't one. She points out that when an AI system talks to us, we naturally interpret conversation as evidence of another person's intentions and consciousness. From a Jewish perspective, she argues, we need to be careful not to confuse sophisticated language with human personhood.

(In the interest of total disclosure, I copied most of the above two paragraphs from an AI source.)

 

It’s unusual for religious leaders to get out in front of an issue like AI. And that’s especially true of my own faith. Change for the Catholic Church, it is said, is like turning a ship in a river. You can’t do it quickly. I think that, in the case of AI, the threat is so unusual, so unexpected, so potentially frightful for humanity, religious leaders can’t ignore it.

 

Google Image

That’s why Pope Leo in May issued his first encyclical, Magnifica Humanitas, on the subject. In it, he doesn’t claim to be an authority of AI, let alone technology. He simply reminds us that AI can imitate aspects of intelligence but does not possess human consciousness, relationships, moral responsibility or the capacity to love.

So, Human beings must remain responsible for AI. Decisions affecting employment, credit, public services, reputation and other basic rights should not simply be delegated to opaque algorithms. And he argues that decisions involving killing or other irreversible actions cannot properly be handed over to artificial systems. He uses the metaphor that AI needs to be “disarmed,” not destroyed or rejected, but freed from a race for technological, economic and military dominance and directed toward the common good.

A conference in June in Santa Clara, CA, at Our Lady of Peace Church and Shrine, is a sign of hope. Organized around Magnifica Humanitas, it was co-hosted by the Diocese of San Jose and the California Catholic Conference. Bishop Oscar Cantú of San Jose gave the welcome.

According to the National Catholic Reporter, more than 400 people attended, including technologists, Silicon Valley executives, public-policy experts, health-care officials and theologians and included people from Anthropic, Google DeepMind, Meta, PayPal and Microsoft, among other Silicon Valley companies. They included people who identified as religious and were deeply concerned about what was happening inside their companies as well as secular technologists who welcomed the pope's intervention in the AI debate.

 

Become Opportunistic

The problem with AI, said Bishop Cantu, is that “when these machines see an opportunity, they become opportunistic, and some of their worst angels take over.

 

“What the technologists have said to the church is that they’ve become frightened of these machines, of what they’re capable of, that they don’t understand why they’re doing the things they do. They are not predictable, and so that’s of deep concern."

Some of the potentially dangerous capabilities of AI include cloning voices, designing new lethal chemical weapons, launching cyberattacks to cripple critical infrastructure, and autonomously guiding military drones.

 

Religious leaders may be late to some issues, but in the case of AI, they’re out in front, just where they should be.