CNA Staff, Jan 6, 2025 / 06:00 am
While humanity continues to debate its uses, merits, dangers, opportunities, and ethics, the development and use of artificial intelligence (AI) continues apace, making it difficult for anyone to guess just how it might shape the world in 2025. Some of the biggest AI systems, including those created by OpenAI, are already said to be close to achieving artificial general intelligence, or AGI — essentially, they’re almost as smart as humans, and their creators are far from done.
The Catholic Church has been no passive observer of the AI revolution; it’s actively shaping its development. By combining cutting-edge technology with the wisdom of the Church, Catholics are working to try to ensure that AI serves humanity and promotes the common good. Most notably, Pope Francis himself has spoken frequently about the importance of developing and using AI in an ethical manner.
So how might Catholics use and shape AI in the coming year, and can we predict where “Catholic AI” might go next?
CNA spoke with several leading Catholic experts on artificial intelligence to see where the technology might be headed in 2025: Father Michael Baggot, LC, a bioethics professor at the Pontifical Athenaeum Regina Apostolorum in Rome; Father Philip Larrey, a professor of philosophy at Boston College who previously served as dean of the philosophy department at the Pontifical Lateran University in Rome; and Matthew Harvey Sanders, an entrepreneur and founder of Longbeard, a company that is developing several high-profile AI tools for Catholic applications.
The end of the world as we know it?
It’s the elephant in the room, so we may as well address it first: Will AI take a sinister turn — insert nightmarish sci-fi movie plot here — and destroy us all in 2025?
The question may sound glib, but it’s not as far-fetched a scenario as you may think — at least according to computer scientist Geoffrey Hinton, often described as the “godfather” of AI, who recently prompted frantic headlines the world over when he opined that there is a “10% to 20%” chance that AI will lead to human extinction within the next 30 years.
For his part, Larrey said he has thought a lot about this possibility of AI human extinction and said he is inclined to take Hinton’s assessment seriously. And in fact, Larrey signed on to an open letter in 2023, joining various experts and pioneers in AI in calling for a six-month moratorium on the development of more powerful AI systems than those already released at the time.
How could this extinction come about? Well, Larrey said, what if an AI with malicious intent gets ahold of nuclear codes? Or decides it wants to trick unsuspecting humans into synthesizing and releasing a deadly virus? Or, even without malicious intent, what if an AI sets a doomsday scenario in motion by accident, or as a result of a hack by bad human actors?
For Larrey, who has spoken and thought a lot about how to ensure AI are imbued with ethics, he said a phrase uttered by Elon Musk years ago rings true to him: “We have to make sure AIs consider us an interesting part of the universe.”
And a hopefully effective way to do that, Larrey said, is for human beings to treat one another with greater respect and reverence so the AI can observe that behavior and learn from it.
Human behavior is one of the strongest arguments AI can come up with when considering how it should treat humans, he continued, and if humans do not treat one another with respect and dignity, AIs may not see them as deserving of special consideration.
In addition, Larrey said he believes that absent a moratorium on further AI development — which looks unlikely at this point to happen — Catholics should continue to promote dialogue between ethicists and AI developers. Larrey said he has spoken frequently with Sam Altman, the CEO of OpenAI and a leader in the space, to encourage him to prioritize human flourishing in his company’s further development of AI.
“Right now, we don’t have a universal code of ethics or anything about AI. I think that the Catholic Church can provide guidance until we get one. I think that’s what we’re trying to do now,” Larrey said.
“If we get this wrong, it’ll be too late to correct it. The damage will already have been done.”
(Story continues below)
Catholic knowledge and teachings for all
So what about some of the positive opportunities that AI presents for the Church?
One application for AI that has been growing in leaps and bounds — and is certain to make even more leaps in 2025 — is the ability of AI to draw on and distribute Catholic teachings and knowledge in previously unimaginable ways.
The world has already been graced with tools like CatéGPT, a chatbot designed to provide accurate and thorough answers to questions about Catholic teaching by drawing on authoritative documents. Users have also found success with secular AI tools such as Google’s NotebookLM, which has the ability to generate realistic-sounding podcasts after being fed source material by users and can be used to distill and explain lengthy Catholic documents, such as the final document of the Synod on Synodality.
But Longbeard, the company spearheaded by Sanders, is poised to make an even bigger breakthrough in this space in 2025. It already offers Magisterium AI, a generative AI product focused on making Church teaching and Catholic insight more accessible by ensuring generated responses are grounded in Catholic sources. And Vulgate AI, a powerful library platform and another Longbeard product, is already enabling the digitization and preservation of library holdings at Catholic universities, making these resources useful and accessible to a global audience.
But Sanders told CNA his company has its sights set even higher for 2025.
Sanders’ team is developing Ephrem, the world’s first Catholic language model, which will be trained on the entire corpus of “Catholic knowledge” — a task easier said than done, since a good chunk of the world’s Catholic knowledge is in book form and safely stored away in libraries or at the Vatican.
Sanders told CNA that by March, the company hopes to have a state-of-the-art robotic system up and running in Rome called the Alexandria Digitization Hub, which will set about scanning thousands of ancient Catholic books and texts, many of which have never even been translated before, let alone digitized.
Once it has digital copies of every piece of “Catholic knowledge” ever produced, it will complete Ephrem’s “training” on it and use it to power Longbeard’s AI tools.
The advantage here, Sanders said, is that the AI will have a truly Catholic foundation for its operation rather than relying on secular AI foundations such as ChatGPT, which are trained on diverse datasets with potentially conflicting values and thus have ingested lots of non-Catholic information, meaning they sometimes display dubious answers when it comes to Catholic topics.
Longbeard will make Ephrem available for Catholics to use — to create apps, products, and other tools — through its API (application programming interface), Sanders said, meaning vast quantities of Catholic teaching documents and books — 80,000 so far, with lots more to come — will be accessible to anyone with an internet connection and available for Catholics to use in creative new ways.
“Any kind of Catholic ministry or organization which is looking to leverage AI, but they want to make sure the AI is faithful, they’ll be able to actually build on top of us,” he explained.
AI agents: Assisting priests, Catholic educators, and more
A much-hyped application for AI in the secular world is the arrival of AI agents — applications that can be called upon by users to perform tasks such as ordering food, booking travel, and more.
Sanders said the company hopes that in the new year, Catholics will be able to use the platforms of Magisterium and Vulgate to create useful AI agents that can perform tasks to serve the Church.
A prayer app could create an AI agent that users could use to look up information on a saint mentioned in a prayer session, for example. Or, a teacher could use it to create a personalized lesson or tutoring plan for a student, making the agent an “active partner in the process of doing theology.”
“What we’re trying to do is provide Catholics like teachers and priests access to AI-powered tools, basically to help them accelerate the work that they’re doing, whether it be preparing a homily or preparing a lesson for students and so on,” Sanders said.
Of course, there’s only so much that a nonhuman can do when it comes to assisting Catholics, especially clergy — naturally, AIs will never be able to hear confessions, for example, or celebrate Mass or any other Catholic sacrament.
But people bullish on AI have differing ideas of just how involved AIs can or should be in the Church’s sacramental life. Notably, an AI-powered art exhibit in Switzerland caused a stir in November when numerous news reports trumpeted that a new artificially intelligent “Jesus” had begun taking people’s confessions.
Despite being placed in the confessional booth, it later became clear the AI installation was intended for conversations, not confessions (which can only be performed by a priest or bishop, and never in a virtual setting.) A theologian at the Swiss parish commented, however, that he believes AI can help with the “pastoral” side of Catholic life.
For his part, Sanders said his company’s work is about “enhancing human human capabilities” in order to help the Church, including priests, and is not intended to replace anyone.
“Those of us who understand the sacramental nature of the Church know that it’s a very much an embodied experience. You need to physically be there. And so there is no realm of possibility in which an AI is going to be able to be part of the sacramental economy. It’s just not going to happen.”
AIs are(n’t) people, too
To go along with the caveat of not letting AIs too far into the Catholic sacramental setting, another point that all three experts emphasized was the importance of not anthropomorphizing AI — in other words, treating it like the tool it is and not as a replacement for true human interaction.
Sadly, numerous reports of people latching too strongly onto humanlike AIs grabbed headlines in 2024, most infamously the story of a 14-year-old from Florida who developed an unhealthy attachment to his chatbot “girlfriend” and later killed himself.
Baggot emphasized that AIs, while being “tremendous tools that can be put at the service of human dignity and human flourishing,” there is cause for concern about the spread of “ever more humanlike AI companions” that lonely people are increasingly turning to.
“AI and other technology should also be an aid and support to our most important relationships with God and neighbor. Such technologies should never seek to replace or detract from these key relationships for human flourishing,” Baggot told CNA.
“Unfortunately, various chatbot companions that claim to be friends or even lovers often blur this important distinction. They may be made with the good intentions of relieving loneliness, but they often distract from authentic interpersonal contact. They can also exploit human vulnerabilities and manipulate humans in unhealthy ways.”
While some recognition of humanlike qualities in machines is natural, Baggot noted it is vital to remember that unlike machines, humans not only have free will but also a unique capacity for insight into meaning and abstract thought.
“We miss out on precious opportunities to encounter God’s love through an empathic human encounter when we have false expectations for our AI tools,” Baggot said.
“AI systems cannot understand and share our emotions because they are not sentient human inventions. They may sometimes appear human and mimic human emotions but lack an inner conscious life. AI systems can provide helpful information about sins and virtues, but it does not know what it is like to struggle through temptation and find liberation through grace.”
Larrey echoed Baggot by saying it is important to remember and to emphasize that AI systems, despite their impressive capabilities, do not understand meaning and are merely manipulating symbols. They can be used to accomplish certain tasks to great effect, but one must be cautious not to think of them as truly human — AIs are not capable of original thought but rather operate by recognizing patterns in their training data and applying them to new scenarios. In addition, he reiterated that while an AI can mimic emotions, it does not actually experience them.
ChatGPT, for example, “simulates understanding, and it’s getting better and better at this, but that doesn’t mean the same thing as understanding,” Larrey said.
Nuclear energy, medical breakthroughs
The training and use of AI requires significantly more computational power — and thus more electricity — than conventional computing, a problem that is already vexing those concerned about the potential environmental impacts.
The revival of nuclear power plants specifically to power AI made headlines in 2024 as companies like Microsoft and Amazon made moves to get nuclear plants online to power their latest power-hungry AI projects, with more such nuclear projects expected in 2025 and beyond.
For his part, Sanders said that because the development of AI has the possibility to solve many societal problems, he believes that the development of AI should continue, and therefore a large source of energy is required. Nuclear energy for AI is likely a good solution, he said, as it is “the safest way to achieve energy abundance at scale.”
Another interesting application for AI that is relevant to the Church is in the field of health care. Baggot said he sees significant potential for AI to improve health care, particularly in making it more predictive, preventative, and personalized — for example, AI systems can analyze medical images such as X-rays, MRIs, and ultrasounds more quickly and accurately, and AI-powered sensors could detect medical problems or worrying mental health indicators in users.
He emphasized, though, that AI should be a tool used to enhance the human element of health care, not replace it; he spoke of the importance of the doctor-patient relationship and the need for human empathy, which AI cannot provide.
“These digital interactions should always direct the user toward interpersonal encounters,” Baggot said.
