Magisterium AI: A Game Changer for the Church?

The new technology holds plenty of promise, but some concerns exist.

Banner for Magisterium AI.
Banner for Magisterium AI. (photo: Courtesy photo / Magisterium AI)

VATICAN CITY — U.S. entrepreneur Matthew Sanders says a new artificial intelligence project that his company has founded “could be a game changer” for the Catholic Church. 

“Our core objective is to make Catholic insight and Church teaching available to anyone in the world, on any device, in their native language,” said Sanders, the founder and CEO of Longbeard, a U.S.-based software company that has created Magisterium AI, a growing dataset of Church documents and algorithms that aims to make the Church’s teaching more accessible than ever. 

Currently experimental, Sanders told the Register that Magisterium AI is aimed primarily at formators and teachers of the faith, helping priests to enrich their homilies, facilitate catechism classes and to assist parents in catechizing their children. 

Among its services, Magisterium AI’s creators say it can answer “any questions” on Church teaching, practices or other topics, helping to “explain complex theological, philosophical, and historical concepts in simple, understandable language.” 

It can also “provide contextual information” on the Church’s history, “helping users understand why the Church teaches what it does,” as well as generate theological reflections, summarize Church documents, and be used to create educational resources. 

Questions and answers can also be given in multiple languages, including French, Spanish, Portuguese and German, and Sanders said they have plans to create a user community where people can “upvote new features and provide suggestions on documents to train our AI on.” 

Their company’s most ambitious plan is to incorporate the entire library of the Pontifical Oriental Institute by digitizing its contents and then adding the documents to their database, enabling the AI to “train on them” and make such resources widely available. It would, Sanders said, be similar to a “Google-Books platform for accessing theological and philosophical works.” They are also teaming up with the Massachusetts Institute of Technology to improve the service still further.

“It’s an enormous opportunity to make available the wisdom and knowledge contained in this unique to everyone,” said Sanders. “We plan to partner with additional pontifical universities in the near future.”

Legionary Father Michael Baggot, professor of bioethics at the Pontifical Regina Apostolorum Athenaeum in Rome and a collaborator on the project, told the Register that he believed “many will find the question-and-answer format engaging and appreciate the concise responses.”

Another adviser on the project, Andreas Widmer, director of the Arthur & Carlyse Ciocca Center for Principled Entrepreneurship at The Catholic University of America’s Busch School of Business, told the Register that he was “very happy” to join Sanders in supporting such a “worthwhile effort” and added that the project’s “possibilities are endless.”

A former Swiss Guard, Widmer said he had recently asked the program to create a syllabus on Catholic Social Doctrine to teach to various ages. 

“It was a lot of fun to see how it adapted that for the different age groups,” he said. “While the topics remain the same, the examples it gave to explain the different principles were amazingly appropriate.” 


More Work to Be Done

But Sanders conceded that a lot more work needs to be done on Magisterium AI before it’s fully workable, and one of the main focal points is to create a wholly comprehensive knowledge base on magisterial teaching. The programmers are currently uploading 20 documents a day, Sanders said, but added: “We still have a long way to go.”  

One such area requiring more input is canon law. Sanders said his company plans to insert commentaries on the Code of Canon Law and upload the canons of the Eastern Church. “It’s not useful yet,” he said, “but it will be incredibly positive.”

In addition, he said the way artificial intelligence is executed also needs improving, especially in the area of “nuance.” The way it “preps homilies can be a little rough,” Sanders said, adding it is “much better at putting together magisterial teaching conducive to a homily.” 

A further question is how it will deal with increasing disputes over the magisterium, or possible changes to the Catechism of the Catholic Church. 

Sanders stressed that he and his team “carefully monitor the websites of the Dicastery for the Doctrine of the Faith,” looking out for “any new decrees or doctrinal statements” and prioritizing them for uploading into their knowledge base. He added that “a document released 60 years ago may be authoritative, but if a newer one comes out and updates a canon or doctrinal understanding, that document will take precedence.” 

“It does it okay some of the time but doesn’t do it as well as we’d want it to, and it’s an area we’re wanting to improve,” he said. “In the next version of AI, it’ll get better at that.” 

The more general concerns around AI are well documented, especially when it comes to academia and research. One of them regards the accuracy of the data and that AI can be prone to mistakes, eventually ending up in research articles. 

But the creators behind Magisterium AI say it works slightly differently to other larger AI programs such as ChatGPT and Google Bard, which have had a mixed reception in academia, in that the information is limited to authentic Church data.

Father Baggot said that unlike more “general generative AI programs,” Magisterium AI draws on a “restricted database of official Church documents” which makes it “much less prone to give false or misleading responses based on unreliable sources.” By including references and links to official Church documents, he said the program enables users to read the original sources. 

However, he cautioned that “since any generative AI system can ‘hallucinate’ [an AI term for creating false information], users should always consult the original documents to avoid confusion.” 

AI is often looked down upon in academia, principally because it tends to work against the requirement for students to have an interior grasp of the truth of an issue and an apprehension of what the words of a particular subject mean. It can also deter engaging in critical thinking, and generally homogenize thought, thereby actually lowering standards in education. Furthermore, AI can be easily used to plagiarize (and be harder to detect),and deter the user from studying the actual documents themselves. 

One academic, who preferred to speak anonymously, told the Register that he feared Magisterium AI could make preaching sound “bureaucratic,” producing boiler-plate homilies without drawing on the preacher’s own lived experience, and that in general research, it would “inevitably bias modern documents” as more ancient ones, including many in Latin, won’t be on the database. 


Answering AI Concerns

Jesuit Father David Nazar, rector of the Pontifical Oriental Institute, said AI is “surely a tricky area” but he added that the “concerns are not new,” especially over plagiarism, and he referred to several world leaders who plagiarized before the days of computerization and digitization.

He said that at the institute, professors “who know their field usually spot an uncredited text in a thesis,” as well as know their students well and their capabilities. The students themselves must defend their thoughts in oral defenses, he added, and so using AI material in their work would defeat their purpose for being a student there. 

As for other concerns regarding AI, Father Nazar said users need to be “encouraged to use the tool respectfully and creatively, in other words, to use it for the good,” and that aids to preaching, both old and new, have usually always improved it. “People know a good homily,” he said. “Boiler-plate homilies will only empty the pews, something no one particularly wants.” 

Widmer downplayed fears over such technology, saying that AI is really what are called “Large Language Models” (LLMs) rather than anything sinister like “HAL in Space Odyssey,” and that its effectiveness is really only as good as the incoming data. “As the saying goes, the output is only going to be as good as the input,” he said, but he believes that it will make various aspects of research work within the Church quicker and more efficient.

What it needs, Widmer said, are individuals who can make sense of the data and use “their agency to apply it, convey it to others” and thereby allow people “to focus more on what we do best: Be creators.” 

He added, “It’s what we humans do, because we are made in the image and likeness of the Creator. I don't see that ever going away.”