Not all Catholic AI bots are creepy: Some new tools for learning about the faith

 

CNA Staff, May 6, 2024 / 06:00 am (CNA).

If you have a question about a teaching of the Catholic Church in 2024, where do you go for a solid answer? You can crack open the catechism yourself, ask a trusted personal source like a priest or theologian, or you can delve into that famously infallible repository of knowledge — Google.

A friendly arms race of sorts has arisen among Catholics around the world to provide Catholics with another option, however — one based around artificial intelligence (AI). In the past year or so, several online AI tools have been released that generate authoritative-sounding answers about Catholic teaching based on users’ questions.

You may also have heard about one recent and unfortunate misfire: an AI “priest” created and unveiled last week by the California-based apologetics apostolate Catholic Answers, which was criticized by some users for its video game-like priestly avatar.

Moreover, at least one user managed to goad the character into providing “absolution,” prompting a statement from the apostolate in which it promised to replace the priest character with a lay character named “Justin.” Catholic Answers’ leaders have expressed optimism about the project, despite the initial public setback.

Meanwhile, Catholics looking for AI-powered answers have other, avatar-less options, like CatéGPT, an artificial intelligence chatbot designed to provide accurate and thorough answers to questions about Catholic teaching by drawing on authoritative documents.

Nicolas Torcheboeuf, a 31-year-old Swiss engineer and a Catholic, developed CatéGPT in his spare time and launched it in the late spring of 2023. (“Caté” is French for catechism, and the name is also a play on the name of the groundbreaking secular chatbot “ChatGPT.”)

The simple online tool accepts a user’s question related to the Church’s teaching — “Why is baptism necessary?” for example — and provides a succinct summary of the answer, citing sources and categorizing the sources by type, making distinctions between encyclicals, Scripture, canon law, various writings of popes and Church Fathers, and other authoritative Catholic sources.

This concept might sound familiar — Torcheboeuf concocted the idea for CatéGPT around the same time that the similar U.S.-based Magisterium AI made its debut. Also in the online ether is Catholic.chat, an interactive platform that allows users to engage with the catechism in a natural, conversational format. The similarities between the various projects, Torcheboeuf said, “goes to show that our intuition was right and meets a real need.”

Despite the free tool’s impressive ability to summarize answers to complex questions about the Church’s teaching, Torcheboeuf said his invention primarily aims to encourage Catholics to read the relevant Church documents for themselves.

“In addition to providing clear answers, [CatéGPT provides] a list of reference documents to encourage the user to read them,” he explained.

“The main aim is above all to invite the user to rediscover the wealth of documents that the Church has produced over the course of its existence, and which constitute a formidable heritage for understanding the world in which we live,” Torcheboeuf said.

Torcheboeuf presented his invention, which is available in English, French, Spanish, Italian, German, Polish, and Portuguese, at the 58th World Day of Social Communications at the Vatican in February. He said the project was initially “born out of the craze surrounding ChatGPT,” OpenAI’s powerful chatbot that burst onto the scene in late 2022.

“I was quickly impressed by the power of artificial intelligence and the number of tools that could be developed using this technology,” he said.

Tinkering with the tech, he started creating chatbots that used customized databases to provide answers within specific fields.

“That’s when I came up with the idea of creating a chatbot that uses the teachings of the Catholic catechism and the texts of the magisterium: These texts exist completely freely and don’t change too much over time, which means that the answers are reliable and stable over time,” he noted.

The idea for CatéGPT, which is kept afloat entirely by donations, didn’t come out of nowhere — Torcheboeuf said his motivation for the project “corresponds to a concern I’ve had for a long time.” He said the area where he lives in Switzerland, while economically prosperous, lacks a vibrant practice of the Catholic faith.

“I’ve noticed that young Catholics today have a fairly low level of education; we’re often called upon to debate fairly complex social issues, and unfortunately we don’t have enough intellectual knowledge to do so properly,” Torcheboeuf said.

“Before trying to reform everything, we need to rediscover the fundamental texts of the Church. When we read these texts, we realize that many of the questions we ask ourselves are answered in encyclicals and catechisms.”

Torcheboeuf’s tool isn’t infallible, of course — no AI is. But the fact that CatéGPT makes use of publicly available documents on the Vatican website means that its curated sources are virtually guaranteed to be solid, and also that the tool’s database is far less complex than a massive AI like ChatGPT, which might be called on to opine on any topic imaginable.

Still, if the idea of asking an AI for guidance on the Church’s teaching makes you wary, you’re not alone — Catholic Answers’ AI, though well-intentioned, was less than favorably received.

Catholic Answers’ “Father Justin” — clearly an attempt to give a Catholic AI a more pastoral, human face — may have misfired, at least in its initial form. But the idea of making AI more pastoral is one that Torcheboeuf endorses. After all, he said, AI in its current form can be great as a training tool, but “it won’t be able to fully assist the Church in the way that priests, religious, or people fully invested in the Church can.” He said he is in the process of integrating video clips from “Catholic influencers” into CatéGPT’s answers in an attempt to “put a human face behind the theoretical answers.”

The Church under Pope Francis has been engaging with the idea of AI long before the release of ChatGPT. The pontiff, on numerous occasions, has called for the ethical use of the technology and is scheduled to speak at the G7 summit in June in Italy about the ethics of artificial intelligence, amid much talk in the wider world about the threats that AI could pose to humanity.

The explosion of generative AI tools and applications in recent years constitutes a revolution, Torcheboeuf said — and like any revolution, “it can be dangerous.”

Still, Torcheboeuf is quick to point out that “artificial intelligence is only intelligent if there’s a real human intelligence behind it.”

“I think that rather than being afraid of this technology, let’s try to be a player in this field and exploit its positive aspects. Right now, this technology is in full expansion, and there are places [for it] to be taken while remaining careful of course.”

Torcheboeuf said he expects that CatéGPT users will be surprised by the answers they get, in the sense that they will realize, perhaps for the first time, that “the Church has already asked itself most of the contemporary questions and answered them, with great wisdom and coherence.”

“Little by little, we hope that with the help of the references given at the end of the answers, they will read the fundamental texts of the Church and come to understand and reappropriate this heritage,” he said.