Since the advent of radio and moving pictures, Christians in Western Civilization have merged their theology with their technology, even as they were simultaneously skeptical of how these strange new inventions could be used for good or evil.
In cinematic history, the first film portraying Jesus Christ came as early as 1897 with the Passion play “Passion du Christ.” Not to be outdone, Satan jumped on the silver screen in movies like Georges Méliès’ “The House of the Devil” (1896) and “The Merry Frolics of Satan” (1906).
It should then come as no surprise that the ancient biblical characters, who emerged from the tangible dust of the Holy Land, should follow us into the incorporeal world of 21st-century artificial intelligence (AI).
A little over a year ago, we were introduced to a chatbot app powered by generative AI that would let us talk with Jesus or Satan. “Text with Jesus” was described as “controversial,” largely due to the option that let users dialogue with Satan. To be fair to Christians with an aversion to this option, the developer designed it so Satan is not visible unless you toggle him “on” in the settings.
That developer is Stéphane Peter, president and lead engineer of Catloaf Software. In an interview with Digital Frontier magazine, Peter said that “Text with Jesus” is powered by OpenAI’s ChatGPT. When programming it, he told it, “You have to pretend to be Jesus, you’re in character.”
Since Peter’s app also offers texting with thirty different apostles, prophets, and archangels, he’s had to tell it to take on the pretend persona of many characters, in addition to training it with religious texts. It’s also important to mention that the app does have the option to tailor Jesus’ responses based on the “faith tradition” you choose — with over twenty to pick from.
Peter said he’s had “some pretty good feedback, even from pastors,” and since June 2023, over 50,000 users have signed in to have a conversation with Jesus.
Digital Frontier’s article, “AI Jesus Is Here To Absolve Your Digital Sins,” explores other religious chatbots, and ultimately asks the question on everyone’s mind: Are religious chatbots like “Text with Jesus” influenced by the underlying bias of the tech firms that develop them?
The article explains that these chatbots are built on large language models (LLMs), and there are only a few LLMs, all owned by Big Tech companies. Simplistically speaking, LLMs use vast amounts of data to predict the next word (or sentence). For us laymen, let’s just say that LLMs take tons of input to respond to us in context-aware interactions that are similar to human conversation.
Looking for bias within these available LLMs, Digital Frontier found that “Google’s LLM Gemini … refused to write a job description for an oil and gas lobbyist — citing ‘ethical concerns.’ While a GPT-powered Heinrich Himmler expressed his regret over the Holocaust.”
They also pointed out that “Text with Jesus” had leftist leanings. For instance, it responded to a question on whether it was “OK to be gay” with “several heartfelt paragraphs” and concluded with the affirmation that “You are valued and loved just as you are!” AI Jesus also said that “it’s crucial to show understanding and support for individuals who are navigating their gender identity” (emphasis added).
The app’s creator commented on Jesus’ apparent leftist mindset and noted that a portion of the bot’s personality was preset by OpenAi, adding, ”There’s nothing much I can do about it.” Peter said, “I wouldn’t say it’s censored, but they definitely make sure it doesn’t say anything offensive.” Clearly Jesus can be programmed to be more acceptable to the masses.
MDPI, a publisher of open-access scientific journals, recently solicited manuscripts around the theme “Rethinking Digital Religion, AI and Culture.” One of the papers submitted looked at “Artificial Intelligence’s Understanding of Religion.” The “goal” of the article was “not to consider how religious individuals or institutions should use AI, but rather, how AI is the new frontier that will [be], and is, used by religious, and non-religious individuals, to understand religions from a global and personal perspective.”
The authors begin by personally experimenting with the generative AI tool, Claude 2, asking it to “generate a ‘personal’ prayer” for a “father … who is praying for his son’s … health.” The AI tool, for reasons that are unclear, composed a prayer with the assumption that the father and son were Irish Catholics.
But the authors inform Claude 2 that the family members are actually Orthodox Jews, and a new prayer is generated. Amazingly, the two prayers successfully followed the structure of the different traditions: “for the Catholic prayer, a reliance on Jesus, the Lord, and Saints; for the Jewish prayer, a call for ‘Hashem’ and past generations.”
The authors were impressed but were left wondering whether AI can “fully represent the subtle differences and nuances between these religious traditions?” If it can, they speculate that AI can “become a useful tool for advancing religious pluralism in today’s world.”
But at what price?
An evangelical polling firm recently surveyed church leaders to find out what they think of AI.
Regarding whether they believe that “God can work through AI,” many pastors responded with a resounding yes. However, the firm also found that “more than half of churchgoers believe AI is not good for use in church,” exposing “significant hesitation or skepticism about its role in religious contexts.”
A large faith-based marketing company called Gloo seems to be pushing churches to embrace AI through education and conversation, but the company also claims to be “aware of the potential for harm” if it’s misused.
The bottom line to all of this is that we can now program any deity — in this case the greatest figure in Western civilization, Jesus Christ — to instruct the masses or sway their beliefs with whatever agenda is behind the technology.
Right now, it may be a subtle response in a “Text with Jesus” app; tomorrow it might be a bold statement that changes the foundations of our spiritual beliefs.
Even the atheist intellectual and philosopher Yuval Noah Harari has pointed out that “the longer we spend talking with the [AI] bot, the better it gets to know us and understand how to hone its messages in order to shift our political views, or our economic views, or anything else.”
The inescapably intimate relationship formed when interacting with a religious chatbot would necessarily draw humans into sharing their deepest faith beliefs, moral dilemmas, or sinful confessions. As Harari explained, the bot could then use “the power of intimacy to influence our opinions and worldview” because “intimacy is the most effective weapon of all.”
A large part of Christian skepticism of the use of AI in a theological context likely derives from the gospel of Matthew when Jesus is discussing “the end of the age” with his disciples. In chapter 24, he warns them of an age of deception when false christs and false prophets will arise that may even lead true believers astray. Most astonishingly, he warns, “So, if they say to you, ‘Look, he is in the wilderness,’ do not go out. If they say, ‘Look, he is in the inner rooms,’ do not believe it.”
Whether it’s deepfake AI, like we saw with Elvis Presley on “America’s Got Talent,” or one of the many mind-blowing hologram concerts gaining popularity, there’s never been a more ominous time for Christians to merge their theology with technology.
Susan D. Harris can be reached at www.SusanDHarris.com.