They identified the problem: too many questions to answer individually and inability of existing answers to satisfy inquiries. Not only does an AI chatbot directly addresses that problem, it’s precisely the problem AI chatbots are optimized to solve.
We are approximately in the stage when we (humans) discovered radioactive isotopes, or X-rays, both of which are here to stay and were super fun to play with.
I'd be fine with an AI-driven research tool for people to find the human-authored articles CA has, but chatbots and animated priests are no way to do this.
Also, I've listened to CA for years, and I'd be shocked if it takes only eight questions in the course of a show. Eight per hour seems barely right, or even low (unless Jimmy Aiken, bless him, is the apologist on duty).
Agreed, I think there's tremendous value in using AI to parse human input and then select the most relevant articles or excerpts to read and synthesize themselves. There are just too many examples of these chatbots breaking, both intentionally (through bad actors looking for poor questions) or unintentionally (from tweaks to the underlying model or unexpected reactions to new input/training data), to trust them for something so important.
If I'm not mistaken, that's commonly part of the AI development process - use an existing search engine but with a feedback loop to improve the system before release. Not the "some folks around the office played with somebody's AI tool, then we had some other people play with it for a while" technique.
To be fair, I tried it out today and he did give an awesome response (concise even!) on the difference between Thomist and Molinist approaches to predestination. The chatbot potential for getting *good* answers to Catholic questions is one thing— I’m less sold on why that needs to come from an avatar with a backstory.
Agreed. I asked it to explain what Thomists mean when they say that "an oxygen atom exists in a water molecule virtually, but not actually." It nailed the answer in a concise way, and even brought in a relevant scripture verse (from a deuterocanonical book too, which gives me assurance it is truly Catholic). I am pretty impressed with the bot thus far. I just think it should be a text-based research tool built in to their search-box prompt, and not an avatar priest with a soothing ASMR voice, bird-chirping Assisi background, and animated eye-blinking. Oh and it pronounced "Thomists" with a soft "th", which made me laugh out loud.
In 2016 Microsoft released an early AI chatbot onto Twitter, and the Internet quickly accepted the challenge to train that AI to be as racist as possible. It lasted 16 hours.
I appreciate that the Catholic version of that is "how many canonical crimes can we get this chatbot to commit?"
Mr. Sorenson & his team absolutely have caused scandal by inflicting upon the Church "Father Justin," which simulates the sacraments, about which Sorenson is—at least in this interview—completely unconcerned & unapologetic.
Canon law question: Are Sorenson & his team guilty of violating Canon 1379?
"Can. 1379— § 1. The following incur a latae sententiae interdict or, if a cleric, also a latae sententiae suspension:
1° a person who, not being an ordained priest, attempts the liturgical celebration of the Eucharistic Sacrifice;
2° a person who, apart from the case mentioned in can. 1384, though unable to give valid sacramental absolution, attempts to do so, or hears a sacramental confession."
An AI chatbot is by definition not a “person”, so Canon 1379 obviously does not apply.
Frankly speaking, a animated simulation of a sacrament by an obviously fake AI chatbot should be infinitely less scandalous than, say, the simulation of a sacrament by real living people in a movie or stage production…
Methinks the distinction between a stage/screen actor & the chatbot is that an actor is known to be acting, whereas C.A. asked users to accept the chatbot's replies as representing truthfully the Deposit of Faith.
I love Catholic Answers! They recommended The Pillar on their Live show and that is how I found The Pillar. I think we should give them some grace. They are an outstanding apostolate.
"Wrong" in what sense? Like the James Martin, SJ sense that the Church can bless homosexual unions? Or wrong in the Cardinal Hollerich sense that claims Church teaching on homosexuality is "wrong, not based in science"? Or wrong in the Joe Biden sense that someone can be a "practicing Catholic" and support abortion and transgender ideology?
Or "wrong" in that they attempted to apply emerging technology to help people understand our faith...but with mixed, imperfect results?
I will happily give more credence to the complaints here if the correspondents can provide examples of perfect evangelization. Otherwise, I am grateful to the faithful and orthodox Catholics at CA for a valiant, worthy effort even if it has imperfect and mixed results--just like all of our evangelization efforts.
If they'd hired some mildly malicious teenagers to do their beta testing, they'd have found out it was a bad idea before releasing it.
There was another AI designed by the military to run the defense for a building, video cameras and robotic guns. Years of careful work by programmers, made entirely useless by teenaged marines who put paper bags over their heads to trick the cameras.
Kinda like that COVID tracking software they wanted everyone to install on their phones that warned everyone who'd been in proximity to you if you self-reported COVID. Make a false report, prank all your friends, and ruin a few dozen strangers' day.
Mildly malicious teenagers should be a standard testing group.
To your first point, I think we all get that some errors are more egregious, and more damaging, than others. I also think that refusing to give credence to valid criticism, just because there are no perfect examples of evangelization, is a case of confusing "don't let the perfect be the enemy of the good" with "wow, that "Father Justin" thing is going to cause real harm, we should tell them." It would only be irresponsible to NOT tell CA about the harm their project is doing.
You don't see any benefits to AI? We should use it if or helps us to understand the Catholic faith better. See above for another poster who got a very good answer to a question he had.
So while sections of the Internet express concerns about actual demons lurking within the bits and bites of AI, Catholic Answers recruits AI to lead souls in search of a shepherd. This won’t end well.
The Wright Brothers first attempts at flight may not have ended gloriously but I am grateful for the persistent efforts of many to bring greater good from initial failures. If it leads even a few lost souls to make contact with a priest or Catholic parish or simply to sit in the back of the church while Mass is celebrated, I think it has ended well.
I see/hear lots of Catholics complain that "the Church must meet people where they are" and it usually means that the Church must leave people where they are. For better or worse, our society and civilization as a whole, makes extensive use of online tools to make sense of the world. There are innumerable examples of online resources pulling people into sin and apostasy. Why, if you truly care for souls, would you abandon that arena to the Devil and his minions? Why not contest them in every medium across the planet?? What, exactly, is your plan/proposal to reach these souls and lead them to the Light of Christ?
The interviewer here (not sure if it was Ed, JD, maybe Michelle?) did a good job following up; Sorenson really shied away from acknowledging the real problems here and why the use of AI when trying to talk about the faith or matters of doctrine and the priest avatar, in particular, can cause scandal. He did not want to address that directly! I’m surprised that CA didn’t think this through more carefully before rolling it out.
"But the other part of it is that there seems to be a misunderstanding about how AI works, or a general fear about AI. "
How condescending. No one has expressed fear of AI.
"But that’s the only way this gets better. And if we just don’t work on this at all — if we don’t try to learn about and understand AI — then we in the Catholic world fall behind on it."
It's very clear that no one at Catholic Answers really understands how AI chatbots work, even if they spent six months testing it.
Look, the Church doesn't allow confessions to take place over the phone. People can't receive sacraments over the Internet. Does that make the Church anti-technology, or "behind"? Of course not. The appropriate use of technology is fine. Pretending that an AI chatbot is somehow a person is not fine, and it will never be fine.
I worry that, at the point where the chatbot is doing things restricted to priests (like giving absolution) some folks, especially elderly ones, might think there's a human on the other side of that mediocre animation. Issues like that are the main concern of folks, not whether AI should ever be leveraged to help apologetics.
I’m sorry; the very online reaction to this is a disgrace and embarrassment to the Church. People are scandalized that an obviously fake, animated avatar pretends to perform sacraments? Are we three?
People on the internet are routinely confused by Babylon Bee headlines and other satire sites. Those emails pretending to be from Nigerian Princes are sent because there's enough people who follow the instructions to keep a lot of scammers in business.
I promise there are people who'd think that was a valid Confession, which can be a bit more serious than losing all your money. If a priest posted a video on youtube giving general absolution and said nothing about it not being even slightly valid, that priest would likely get in serious trouble.
Just because "AI is here to stay" doesn't mean that we need to have AI solutions looking for non-existent problems.
Completely correct.
They identified the problem: too many questions to answer individually and inability of existing answers to satisfy inquiries. Not only does an AI chatbot directly addresses that problem, it’s precisely the problem AI chatbots are optimized to solve.
We are approximately in the stage when we (humans) discovered radioactive isotopes, or X-rays, both of which are here to stay and were super fun to play with.
"...super fun to play with." Love the perspective.
yeah, and I think we remember what happened to the humans who played with the radioactive substances "for fun" a little too much.
https://en.wikipedia.org/wiki/Marie_Curie#Death
Yes... The mature fruit of the field probably looks like Funranium Labs (once in a while I remember that blog exists and catch up.)
I'd be fine with an AI-driven research tool for people to find the human-authored articles CA has, but chatbots and animated priests are no way to do this.
Also, I've listened to CA for years, and I'd be shocked if it takes only eight questions in the course of a show. Eight per hour seems barely right, or even low (unless Jimmy Aiken, bless him, is the apologist on duty).
Agreed, I think there's tremendous value in using AI to parse human input and then select the most relevant articles or excerpts to read and synthesize themselves. There are just too many examples of these chatbots breaking, both intentionally (through bad actors looking for poor questions) or unintentionally (from tweaks to the underlying model or unexpected reactions to new input/training data), to trust them for something so important.
If I'm not mistaken, that's commonly part of the AI development process - use an existing search engine but with a feedback loop to improve the system before release. Not the "some folks around the office played with somebody's AI tool, then we had some other people play with it for a while" technique.
To be fair, I tried it out today and he did give an awesome response (concise even!) on the difference between Thomist and Molinist approaches to predestination. The chatbot potential for getting *good* answers to Catholic questions is one thing— I’m less sold on why that needs to come from an avatar with a backstory.
They could've/should've had his backstory be obviously false, with a disclaimer coming from him adding that "Father" is an AI chatbot.
That's awesome and a great example of the power of AI.
Agreed. I asked it to explain what Thomists mean when they say that "an oxygen atom exists in a water molecule virtually, but not actually." It nailed the answer in a concise way, and even brought in a relevant scripture verse (from a deuterocanonical book too, which gives me assurance it is truly Catholic). I am pretty impressed with the bot thus far. I just think it should be a text-based research tool built in to their search-box prompt, and not an avatar priest with a soothing ASMR voice, bird-chirping Assisi background, and animated eye-blinking. Oh and it pronounced "Thomists" with a soft "th", which made me laugh out loud.
In 2016 Microsoft released an early AI chatbot onto Twitter, and the Internet quickly accepted the challenge to train that AI to be as racist as possible. It lasted 16 hours.
I appreciate that the Catholic version of that is "how many canonical crimes can we get this chatbot to commit?"
So NOW I'm interested in using it.
This is the primary reason that I'm interested in using it.
Mr. Sorenson & his team absolutely have caused scandal by inflicting upon the Church "Father Justin," which simulates the sacraments, about which Sorenson is—at least in this interview—completely unconcerned & unapologetic.
Canon law question: Are Sorenson & his team guilty of violating Canon 1379?
"Can. 1379— § 1. The following incur a latae sententiae interdict or, if a cleric, also a latae sententiae suspension:
1° a person who, not being an ordained priest, attempts the liturgical celebration of the Eucharistic Sacrifice;
2° a person who, apart from the case mentioned in can. 1384, though unable to give valid sacramental absolution, attempts to do so, or hears a sacramental confession."
https://www.vatican.va/archive/cod-iuris-canonici/eng/documents/cic_lib6-cann1364-1399_en.html
An AI chatbot is by definition not a “person”, so Canon 1379 obviously does not apply.
Frankly speaking, a animated simulation of a sacrament by an obviously fake AI chatbot should be infinitely less scandalous than, say, the simulation of a sacrament by real living people in a movie or stage production…
Methinks the distinction between a stage/screen actor & the chatbot is that an actor is known to be acting, whereas C.A. asked users to accept the chatbot's replies as representing truthfully the Deposit of Faith.
And what if the bot's answer is no worse than you'd get from DEI Ministries or Hypertrad Central?
I love Catholic Answers! They recommended The Pillar on their Live show and that is how I found The Pillar. I think we should give them some grace. They are an outstanding apostolate.
But fraternally correct them if they're wrong. No one should be given a pass just because they are friendly, right?
"Wrong" in what sense? Like the James Martin, SJ sense that the Church can bless homosexual unions? Or wrong in the Cardinal Hollerich sense that claims Church teaching on homosexuality is "wrong, not based in science"? Or wrong in the Joe Biden sense that someone can be a "practicing Catholic" and support abortion and transgender ideology?
Or "wrong" in that they attempted to apply emerging technology to help people understand our faith...but with mixed, imperfect results?
I will happily give more credence to the complaints here if the correspondents can provide examples of perfect evangelization. Otherwise, I am grateful to the faithful and orthodox Catholics at CA for a valiant, worthy effort even if it has imperfect and mixed results--just like all of our evangelization efforts.
If they'd hired some mildly malicious teenagers to do their beta testing, they'd have found out it was a bad idea before releasing it.
There was another AI designed by the military to run the defense for a building, video cameras and robotic guns. Years of careful work by programmers, made entirely useless by teenaged marines who put paper bags over their heads to trick the cameras.
Kinda like that COVID tracking software they wanted everyone to install on their phones that warned everyone who'd been in proximity to you if you self-reported COVID. Make a false report, prank all your friends, and ruin a few dozen strangers' day.
Mildly malicious teenagers should be a standard testing group.
I like that idea. As a high school teacher, I can tell you that motivated teenagers can certainly find ways to stress the system.
To your first point, I think we all get that some errors are more egregious, and more damaging, than others. I also think that refusing to give credence to valid criticism, just because there are no perfect examples of evangelization, is a case of confusing "don't let the perfect be the enemy of the good" with "wow, that "Father Justin" thing is going to cause real harm, we should tell them." It would only be irresponsible to NOT tell CA about the harm their project is doing.
"AI is not going away" until it does. This is not a premise for anything. Gambling and gemology are also not going away (until they do). So what?
"Phrenology is here to stay."
You don't see any benefits to AI? We should use it if or helps us to understand the Catholic faith better. See above for another poster who got a very good answer to a question he had.
Hammers are not going away. That doesn't mean you should use them to persuade people.
So while sections of the Internet express concerns about actual demons lurking within the bits and bites of AI, Catholic Answers recruits AI to lead souls in search of a shepherd. This won’t end well.
The Wright Brothers first attempts at flight may not have ended gloriously but I am grateful for the persistent efforts of many to bring greater good from initial failures. If it leads even a few lost souls to make contact with a priest or Catholic parish or simply to sit in the back of the church while Mass is celebrated, I think it has ended well.
I see/hear lots of Catholics complain that "the Church must meet people where they are" and it usually means that the Church must leave people where they are. For better or worse, our society and civilization as a whole, makes extensive use of online tools to make sense of the world. There are innumerable examples of online resources pulling people into sin and apostasy. Why, if you truly care for souls, would you abandon that arena to the Devil and his minions? Why not contest them in every medium across the planet?? What, exactly, is your plan/proposal to reach these souls and lead them to the Light of Christ?
I would love to see an AI Catholic Avatar called Vichael Morris, who gets really aggressive and gives the most tardy answer possible.
Somebody call Microsoft Office and see if "Clippy" is available. He'd be more popular than "Fr. Justin."
I haven't met "Fr. Justin", but if he'd be less popular than Clippy, then he's realllly bad. That thing was obnoxious.
CA really should know better than to give this AI the persona of a priest in the first place.
Was it a PR stunt? It seems like a good way to get outsized attention.
Why not "invent" a new CA apologist persona. . .leave the priests out of this. They have enough on their plate.
A virtual mishmash of the current CA apologists.
Jimmy AI-kin
Genius!
I suggested A.I.den the Apologist on the Twitters.
The time spent on developing this would be better spent praying for IRL vocations
Save for cloistered religious, are they really mutually exclusive?
Maybe i was being too subtle. why spend time making a fake priest instead of working to get real ones
Because an AI can work 24/7/365 and speak to thousands of people at once, which a human priest cannot do.
A virtual priest that's guaranteed to give incorrect information every so often.
Well, tbf so are real priests...
Guaranteed to give incorrect information?
As I stated in my earlier comment, a bad and terrible idea.
I suggest that Catholic Answers take the money they're spending on AI and direct it towards hiring an additional staff apologist.
The interviewer here (not sure if it was Ed, JD, maybe Michelle?) did a good job following up; Sorenson really shied away from acknowledging the real problems here and why the use of AI when trying to talk about the faith or matters of doctrine and the priest avatar, in particular, can cause scandal. He did not want to address that directly! I’m surprised that CA didn’t think this through more carefully before rolling it out.
If catholic answers didn't roll this out somebody would. That's a guarantee.
// I’m surprised that CA didn’t think this through more carefully before rolling it out. //
Don't be surprised. These are not mature adults.
"But the other part of it is that there seems to be a misunderstanding about how AI works, or a general fear about AI. "
How condescending. No one has expressed fear of AI.
"But that’s the only way this gets better. And if we just don’t work on this at all — if we don’t try to learn about and understand AI — then we in the Catholic world fall behind on it."
It's very clear that no one at Catholic Answers really understands how AI chatbots work, even if they spent six months testing it.
Look, the Church doesn't allow confessions to take place over the phone. People can't receive sacraments over the Internet. Does that make the Church anti-technology, or "behind"? Of course not. The appropriate use of technology is fine. Pretending that an AI chatbot is somehow a person is not fine, and it will never be fine.
Nobody is pretending that an obviously fake animated AI chatbot is a real person…
I worry that, at the point where the chatbot is doing things restricted to priests (like giving absolution) some folks, especially elderly ones, might think there's a human on the other side of that mediocre animation. Issues like that are the main concern of folks, not whether AI should ever be leveraged to help apologetics.
I’m sorry; the very online reaction to this is a disgrace and embarrassment to the Church. People are scandalized that an obviously fake, animated avatar pretends to perform sacraments? Are we three?
People on the internet are routinely confused by Babylon Bee headlines and other satire sites. Those emails pretending to be from Nigerian Princes are sent because there's enough people who follow the instructions to keep a lot of scammers in business.
I promise there are people who'd think that was a valid Confession, which can be a bit more serious than losing all your money. If a priest posted a video on youtube giving general absolution and said nothing about it not being even slightly valid, that priest would likely get in serious trouble.