从教堂到聊天机器人:人工智能如何与宗教融合

从撰写讲道词到模拟与耶稣的对话,宗教领袖和信徒们都在尝试使用人工智能——尽管这并非没有争议。
作者:哈尼·里希特
2026年2月7日

打开网易新闻 查看精彩图片

2024 年,加利福尼亚州瓦列霍市友谊浸信会的牧师贾斯汀·莱斯特为他的教会构建了一个定制的 GPT,该 GPT 使用他的讲道来开发小组材料,并允许其他教会领袖根据这些讲道来构建课程。

莱斯特对以这种方式运用人工智能毫无顾虑。在他看来,使用这些工具对灵性成长、门徒训练和社区发展都至关重要。

“耶稣说过我们会成就更伟大的事,”他说。“我认为(人工智能)是这更伟大事的一部分。”

人工智能正在悄然改变人们的工作、生活和恋爱方式。或许,它渗透到人们的宗教信仰中也只是时间问题。然而,随着信徒和宗教领袖开始将这项技术融入他们的宗教生活——从模拟与耶稣的对话到撰写讲道稿——一些学者和宗教领袖警告说,这项技术可能带来风险和潜在危害。

公开的无神论者 Siraj Raval 说,是孤独和对存在的恐惧让他发现了“TalkToHim”,这是一个人工智能聊天机器人,可以模拟与耶稣的对话。

“我曾有过这样的体验:我感觉自己被一种神圣的存在所倾听,”他谈到这款应用程序时说道。他曾用这款应用程序来寻求关于精神问题的答案,例如如何与内疚共存,如何在看似不可能的情况下原谅他人,以及如何以道德的方式行事。

“它比教科书好多了,”拉瓦尔说道,他经常去爱达荷州的一家非宗派基督教教堂。“它比读圣经还好。”

人工智能的这种应用方式并非仅限于个人层面。去年,瑞士圣彼得教堂与当地一所大学合作,在其告解室安装了一个人工智能耶稣化身,作为一项实验性艺术装置。教堂神学家马可·施密德最惊讶的是,人们对这种体验的认真程度,甚至有人对聊天机器人表示感谢。

“你会在用完电脑后对电脑说‘哦,谢谢你,电脑’吗?”施密德说。“但你看,正因为系统太好用了,人们才会赋予它如此多的个性化和人性化元素。”

休斯顿伊曼努埃尔会堂的拉比乔什·菲克斯勒是 ChatGPT 的早期使用者。2023 年犹太新年期间,这位 41 岁的拉比播放了一段自己讨论人工智能对人类影响的录音,令他的教众震惊不已——而他后来透露,这段布道是由人工智能生成的。

但与其他采用这项技术的人不同,他对它的结果并不完全满意。“那次布道之后,我对这项技术的伦理以及人们对技术的过度关注都深感担忧,”他谈到自己的人工智能实验时说道,此后他再也没有重复过这项实验。主要原因是:聊天机器人给出的一些信息根本不属实。

“聊天机器人引用了一位名叫迈蒙尼德的伟大犹太学者的话,但据我所知,这句话是它编造的,”菲克斯勒说。

从20世纪60年代电视福音布道的兴起,到新冠疫情期间宗教场所广泛采用Zoom等在线通讯工具,科技长期以来一直是宗教创新的驱动力。然而,尽管这些工具主要扩大了现有宗教活动的覆盖范围,人工智能似乎正在重塑人们学习、诠释乃至体验信仰的方式。

牛津主教史蒂文·克罗夫特说:“我认为基督教社群的本质具有独特性,它强调人与人之间的面对面交流,以及深刻的人性关怀。其原因在于基督教信仰对耶稣的理解:上帝在耶稣里成为人。因此,基督教本质上是个人化的。”

克罗夫特的犹豫也引起了其他宗教领袖和学者的共鸣,他们中的许多人表示,对人工智能提供可靠宗教建议的能力缺乏信任。苏黎世大学数字宗教助理教授贝丝·辛格勒回忆起Character.ai的“佛陀”就曾错误地声称佛教有五圣谛,而不是四圣谛。但她担心的不仅仅是这些错误。

辛格勒说:“关于宗教领袖形象的伦理问题一直存在争议,尤其当聊天机器人说出亵渎神明甚至危险的话时。我们已经看到一些具体案例,有人因为与聊天机器人的对话而走上自杀之路。关于这种情况发生的频率,有一些非常可怕的统计数据。”

剑桥大学利弗休姆未来智能中心访问学者雅库布·乔杜里也质疑人工智能是否是传递有效且可追溯的宗教信息的最佳媒介——尤其是在他所信仰的伊斯兰教的背景下,伊斯兰教认为《古兰经》是上帝的直接且未经篡改的圣言。

“如果是由大型语言模型(LLM)将他所学的知识混杂在一起,就能真正传达伊斯兰教义吗 ?”他说道,“这在区分清真、禁忌、推荐、允许、禁止和令人厌恶的事物方面,确实是个大问题。”

菲克斯勒表示,尽管人工智能可以为用户提供探索自身信仰的新方法,但它不太可能取代人们对人际交往的基本需求。

“我认为宗教的作用不是让机器变得更像人,”他说。“宗教的作用是让我们所有人都能成为最像人的人。”

From churches to chatbots: How AI is fusing with religion
From writing sermons to simulating conversations with Jesus, religious leaders and worshippers alike are experimenting with AI — though not without controversy.
By Hani Richter
February 7, 20267:00 PM GMT+8Updated February 7, 2026

REUTERS illustration/Anurag Rao
REUTERS illustration/Anurag Rao
In 2024, Justin Lester, a pastor of Friendship Baptist Church in Vallejo, California, built a custom GPT for his church that uses his sermons to develop small group materials and allows other church leaders to build lessons based on those sermons.
Lester has no qualms about deploying AI in this way. As he sees it, using these tools is important for spiritual growth, discipleship, and community development.
"Jesus said we will do greater things,” he says. “And I think (AI) is part of the greater."
AI is quietly reshaping how people work, live, and love. So perhaps it was only a matter of time before it crept into how they worship, too. But as worshippers and faith leaders alike begin to integrate this technology into their religious lives — from using it to simulate conversations with Jesus to writing sermons — some academics and religious leaders warn about the risks and potential harms it could cause.
Advertisement · Scroll to continue

An avowed atheist, Siraj Raval says it was loneliness and existential dread that led him to finding “TalkToHim,” an AI-powered chatbot that simulates conversations with Jesus.
"I had an experience where I felt listened to by a presence that was divine," he says of the app, which he used to seek answers to his spiritual questions, such as how to live with guilt, forgive when it feels impossible, and to act morally.
"It was better than a textbook,” Raval, who regularly attends a non-denominational Christian church in Idaho, says of the app. “It was better than reading the Bible."
Incorporating AI in this way isn’t just happening at the personal level. Last year, St. Peter's Chapel in Switzerland installed an AI Jesus avatar in its confessional booth as part of an experimental art installation with a local university. What most surprised Marco Schmid, a theologian at the church, was how seriously people took the experience, with some even thanking the chatbot.
"Do you say to your computer when you finish, 'Oh, thank you, computer?’ No,” Schmid says. “But you see how much people personalized and humanized the system because it was so good."
Rabbi Josh Fixler of Congregation Emanu El in Houston was an early adopter of ChatGPT. During the Jewish High Holidays in 2023, the 41-year-old shocked his congregants when he played a recording of himself discussing the impact of AI on humanity — a sermon that he later revealed was AI-generated.
Advertisement · Scroll to continue

But unlike other adopters of the technology, he wasn’t wholly impressed with its output. "I came away from that sermon with real concerns about both the ethics of the technology and also the hyperfocus on the technology," he says of his AI experiment, which he hasn’t replicated since. The main reason: some of what the chatbot came up with simply wasn't true.
"[The chatbot] quoted a great Jewish scholar named Maimonides, but as best I can tell, it made up that quote," Fixler says.
Technology has long driven religious innovation, from the rise of televangelism in the 1960s to places of worship's widespread adoption of online communication tools like Zoom during the COVID-19 pandemic. But while those tools primarily expanded the reach of existing worshipping practices, AI appears to be reshaping how people learn about, interpret, and even experience their faith.
"I think there is something distinctive about the nature of Christian community, which is about being in person and face to face and being deeply human,” Steven Croft, the bishop of Oxford, says. “The reason for that is rooted in Christian faith's understanding that in Jesus, God became a human person. So Christianity is inherently personal."
Advertisement · Scroll to continue

Croft’s hesitancy is shared by other religious leaders and academics, many of whom cite a lack of trust in AI’s ability to provide sound religious advice. Beth Singler, an assistant professor in digital religion at the University of Zurich, recalls an instance when a Character.ai "Buddah" erroneously claimed there were five noble truths in Buddhism, instead of four. But it’s not just inaccuracies that she’s concerned about.
"There are questions about the ethics of representations of religious leaders,” Singler says, especially if the chatbot says something profane or, worse, dangerous. "We've seen specific examples of people being pushed towards suicide by conversations with chatbots. There (are) some really scary statistics about how often that happens.”
Yaqub Chaudhary, a visiting scholar at the Leverhulme Centre for the Future of Intelligence at the University of Cambridge, also questions whether AI is the best medium to deliver valid and attributable religious information — particularly in the context of his faith, Islam, which considers the Quran the direct and unaltered word of God.
"Is that a true communication of the Islamic meaning if it is produced by an LLM, mixing together whatever it has in its training set?” he says. "That is a really huge problem in terms of knowing the halal, the haram, the recommended, the permissible, the impermissible, the disliked."
Advertisement · Scroll to continue

As much as AI may offer users new ways of exploring their beliefs, Fixler says it’s unlikely to replace people's fundamental need for human connection.
"I think that the work of religion is not trying to get machines to be more human,” he says. “The work of religion is trying to get us all to be the most human human."
The Reuters Inside Track newsletter is your essential guide to the biggest events in global sport. Sign up here.

Editing by Yasmeen Serhan and Aurora Ellis
Our Standards: The Thomson Reuters Trust Principles., opens new tab