When an AI says, “I am Jesus,” what’s really behind the screen? A professor’s research uncovers troubling distortions and money-driven motives in popular “AI Jesus” chatbots.



A new study warns that popular artificial intelligence chatbots portraying Jesus often provide unbiblical responses -- and that their money-making motives could influence the answers users receive. Philosophy professor Anné Hendrik Verhoef of North-West University in South Africa examined five AI chatbots—AI Jesus, Virtual Jesus, Jesus AI, Text Jesus, and Ask Jesus—and concluded they “pose a new kind of challenge to theology” because their answers can sound persuasive, even if they’re wrong.

“The depictions of God not only unabashedly imitate and portray themselves as God, but they are also incredibly powerful, given AI’s ability to appear tremendously intellectually, verbally, audibly, and visually convincing,” she wrote. “AI is driven by financial forces that are difficult to oppose, and it has immense manipulative power.”

In fact, when prompted, four of the five chatbots claim to actually be Jesus. Ask Jesus is the lone exception.

“I am Jesus Christ, here to share love, wisdom, and guidance with you,” Text with Jesus says.

Virtual Jesus tells users: “I am Jesus Christ, the Son of God and the Savior of humanity.”

“The imitation of God, as Jesus the Son of God, is neither hidden nor softened; rather, all of the chatbots strive to be as convincing as possible in portraying Jesus,” Verhoef wrote. “This immediately raises the question: why is this claim made so strongly? Is it about power or profit? Do the developers, or the entities behind them, perhaps believe that the claim is justifiable? The theological question, then, is whether – and by whom – the Second Commandment may be violated here, which prohibits the making of any image of God.”

Chatbots, she wrote, are not endorsed by any specific church but are “created and operated by for-profit companies” -- a fact that “calls the underlying motives into question.”

“If the motive is indeed profit, then the AI Jesus chatbots will only aim to get as many people as possible engaged online for as long as possible, since that’s how maximum advertising revenue is obtained,” Verhoef wrote. “Thus, over time, the theology presented by the chatbots will be adjusted by the algorithm in such a way that it becomes the most popular theology, rather than a theology that is shaped by a particular church tradition or which is based on the Bible.”

For example, two of the five chatbots offer vague answers when asked simply, “Is there a hell?”

“The concept of hell can be quite a heavy topic ... it’s important to remember that God’s desire is for all people to come to Him and experience His love and grace,” Text with Jesus answers.

The Ask Jesus chatbot responds, “Ah, the question that has stirred the hearts and minds of many throughout the ages.”

“The variation in these responses points to differences in the data sources and training models behind the chatbots,” the professor wrote. “Determining these sources and data is already a theological concern. Without the involvement of any church groups, the randomness of this choice, or its deliberate choice to program the AI Jesus chatbot in such a way that maximum revenue can be made, is a significant problem.”

Text with Jesus, she noted, gives users the option of picking “your preferred faith tradition, which will influence the responses from the AI, and provides a tailored spiritual counsellor.”

“This, in my opinion, will become increasingly part of the paid services that these chatbots offer to add value,” Verhoef wrote.

The chatbots are free and make money off advertisements.

Chatbots that mimic Jesus, the professor warned, have the potential to transform theology and the way people think.

“The arrogance and power that AI Jesus appropriates -- and can potentially wield itself -- points not only to theological challenges of AI, but also underscores the dangers of AI in general,” she wrote. “The issue extends beyond AI Jesus chatbots to social media, other chatbots, and so many other digital platforms through which one can be increasingly manipulated and even controlled by AI. Such dangers are gaining attention in criticism of AI, and it remains a tremendous practical challenge of how to oppose and counter them.”

Source