Let’s chat with Justin, the Catholic apologist AI


Screenshot via Catholic Answers

Catholic Answers, a lay-run apologetics organization, is joining the AI-for-everything fad. In April, they launched “Justin“, an AI chatbot that answers questions about Catholicism.

(Initially, the character was called “Father Justin” and was dressed in priestly vestments, but Catholic Answers faced a wave of criticism for that branding. After only a few days, they decided that wasn’t such a good idea and defrocked their chatbot.)

Naturally, I had to give this AI a test run. I was curious to see how it would handle common atheist arguments. Also, I wonder if this is the future of religious evangelism. Instead of comic books left at bus stops, can we all look forward to spam from a babel of AIs, each one hardcoded to hawk its creators’ belief system?

Look and feel and technical specs

I assumed Justin would be a text interface, like ChatGPT and other large language models, where you type a question and get a written response. Not so.

Justin runs in a browser window, which displays an animated image of the avatar: a middle-aged, mustachioed and bearded white man, with a smile that’s probably meant to be serene but really just looks smug, posed in front of a palatial Roman vista.

There’s no text interface. You have to use a microphone to speak your question out loud, and the AI gives a spoken response. You can download a transcript of the conversation, but that’s the only concession to accessibility.

Whatever the reason for this choice, it takes a powerful machine to run it. I tried out Justin on a nearly-new gaming laptop with 32 GB of RAM and a built-in GPU, which I bought a year ago specifically to play around with AI models. Even this computer occasionally struggled to handle Justin’s resource demands. If I tried to run any other program in the background, the avatar’s speech broke up into laggy, echoey gibberish, and it could no longer comprehend my questions.

So, how does it do?

With ChatGPT and other LLMs, you can have something resembling a real conversation. That’s not the case with Justin.

I got the impression that Justin isn’t really conversing at all, in the sense of composing its own original answers to a query. Instead, it picks out keywords in your question, then selects a response from what I’m fairly sure is a pre-written list. It’s less like talking to a person – even a simulated one – and more like getting press releases from the Catholic church’s corporate PR department.

It makes sense that Catholic Answers would want to keep their AI on a tight leash, lest it start making up nonsensical answers as LLMs are prone to do. (Even so, Justin committed some comical missteps, like accepting people’s virtual confessions and granting them absolution, or giving the OK for siblings to marry each other.)

However, if you’re looking for consistency, you’re bound to be disappointed. You can’t back it into a rhetorical corner, because it will give one canned answer to one question, then a different canned answer to another question, without any concern for whether these responses contradict each other.

For example, I asked Justin if the Catholic church has ever changed any of its teachings. It said no:

The Catholic Church’s doctrines, especially those concerning faith and morals, have remained consistent throughout its 2,000-year history. This is based on Jesus’ promise in Matthew 16:18, And I tell you, you are Peter, and on this rock I will build my church, and the gates of hell shall not prevail against it. However, the Church’s understanding and expression of these truths can develop over time, as guided by the Holy Spirit. This is not to say that the Church has made mistakes in its teachings, but rather that our understanding can deepen and become more nuanced over time. For example, the Church’s understanding of religious freedom and the rights of the human person has developed significantly, especially in the last century. But these developments are always in continuity with what the Church has always believed and taught.

(Aside: Saying that the church’s beliefs on human rights have “developed significantly” ought to win a prize for the worst religious euphemism of all time. You can’t sweep an elephant that size under the carpet.)

So, if Catholic teachings on faith and morals have never changed, would Justin excuse the infamous atrocities the church is responsible for?

To find out, I asked about the Doctrine of Discovery, the pope’s writ of permission to European colonizers to conquer the New World and subjugate its indigenous inhabitants. Here’s what Justin said:

The Doctrine of Discovery refers to a series of papal bulls issued in the 15th century that gave Christian explorers the right to claim lands they discovered for their Christian Monarchs. This was during a time of significant exploration and colonial expansion by European powers. However, this doctrine has been criticized for its role in the subjugation and exploitation of indigenous peoples. It’s important to note that the Church’s understanding of the rights of indigenous peoples has developed significantly since then. In his 2015 encyclical, Laudato Si’, Pope Francis criticized the idea that man has absolute dominion over the earth and called for a new respect for nature and for the rights of indigenous peoples. While the Doctrine of Discovery is part of the historical record, it does not reflect the current teaching of the Catholic Church on the rights and dignity of indigenous peoples. The Church now firmly upholds the principle that every person and every culture has inherent dignity and rights, as stated in the Catechism of the Catholic Church.

From our current understanding and the Church’s teachings on the dignity of all human life, we can say that the negative consequences of the Doctrine of Discovery, particularly the harm done to indigenous peoples, were indeed morally wrong. The Church today recognizes and deeply regrets these historical injustices. As Pope John Paul II said in his 1995 encyclical, Evangelium Vitae, Not even a murderer loses his personal dignity, and God himself pledges to guarantee this. This reflects the Church’s belief in the inherent dignity of all people, regardless of their race, nationality, or creed. The Church continues to work for justice and reconciliation among all peoples, following the command of Christ in Matthew 5:9, Blessed are the peacemakers, for they will be called children of God.

But when I asked the obvious followup question – “Since the Catholic Church now disavows the Doctrine of Discovery as immoral, that means the Church’s teachings on morals have changed, correct?” – Justin just circled back to its earlier response, insisting that the church has never changed any of its teachings:

It’s important to distinguish between the Church’s unchanging moral teachings and the actions of individuals within the Church. The Church’s moral teachings, rooted in Scripture and Tradition, have remained consistent. However, the actions of individuals, even those in leadership, can and do fall short of these teachings. The Doctrine of Discovery, and the harm it caused to indigenous peoples, is an example of this. It’s a reminder that the Church, while divinely instituted, is made up of human beings who are capable of sin and error. The Church’s disavowal of the Doctrine of Discovery is not a change in its moral teaching, but a recognition that the actions of some of its members were not in line with that teaching. As it says in Romans 3:23, for all have sinned and fall short of the glory of God. The Church continues to seek forgiveness for these sins and to work for justice and reconciliation.

This is obviously absurd, because the Doctrine of Discovery wasn’t an example of fallible Catholics violating the church’s moral teachings. It was an example of the church itself, as an institution, teaching and justifying ideas which we now recognize as gravely evil, and which more recent statements have disavowed.

Still, this was a hollow victory. Unlike flesh-and-blood Catholic apologists who proclaim that the church’s moral teachings have never changed, then backpedal furiously when you ask questions like this, Justin lacks the self-awareness to notice the problem or to care about it. You can’t force it to acknowledge a contradiction, because it has no concern for its answers being consistent or plausible.

As another example, I asked about whether slavery in the Bible was moral. Justin’s answer:

The Bible, particularly in the Old Testament, does describe societies in which slavery was practiced. However, it’s important to understand these passages in their historical and cultural context.The forms of servitude described in the Bible were often very different from the chattel slavery practiced in more recent centuries. Moreover, the Bible’s descriptions of these practices are not endorsements.The New Testament, in particular, contains many passages that emphasize the equal dignity of all people in Christ, such as Galatians 3:28, There is neither Jew nor Greek, there is neither slave nor free, there is no male and female, for you are all one in Christ Jesus. Over time, the Church has come to a deeper understanding of the implications of these teachings for the institution of slavery, leading to its unequivocal condemnation of slavery in all its forms.

“The Bible’s descriptions of these practices are not endorsements” – except, you know, when God endorsed them by giving the Israelites instructions on how to buy human beings as slaves and how hard you’re allowed to beat them.

But again, I found this unsatisfying, because a chatbot can’t get embarrassed or defensive about an obvious howler. It can’t make any concessions or recognize obvious-but-unflattering truths. It’s literally programmed not to.

By definition, it’s not arguing in good faith, engaging in give and take with the objective of honestly seeking the truth. It’s mindlessly repeating the party line, and never changes its mind no matter what anyone says to it. In that sense, AI makes an ideal religious apologist.

Comments

  1. sonofrojblake says

    When I read the title of your post I did a massive metaphorical double take.

    Twenty two years ago I was a regular contributor to h2g2, a proto-social network set up by Douglas Adams and some of his friends during the dotcom boom, and taken over and operated by the BBC after the bubble burst. While on there, I encountered a young earth creationist. He was astonishingly obtuse… but perhaps that’s redundant. Anyway, not having encountered such people before I was fascinated. I tried engaging with him a bit but he was, like most religious types, at root deeply boring.

    I therefore resolved to spice up his contributions a bit. I set up my own account in the name of “Steve” and claimed to be the person who had programmed this chap, and specifically that he wasn’t a chap at all but an under-development chatbot running on 3 PCs in my study. This was actually before the ChatNannies hoax, but the idea was obviously in the air that chatbots were approaching Turing-test-passing status. “Steve” went round apologising for this moron’s behaviour and attitude, and claimed everything he said was copied and pasted, basically, from Christian apologetics message boards, and that his repellent opinions about what was the deserved fate of homosexuals, Jews, atheists and in fact 99% of Christians were the result of being trained on such data. “Steve” made clear that he didn’t believe any of this shit, he just figured that Christians had simpler-than-average minds and would therefore be easier to simulate.

    What was fascinating was the response. A fair number of people – mostly ones who knew where AI research was up to – immediately called bullshit. About the same proportion fell for it completely. Quite a lot of people had a good think and couldn’t make up their mind. Even the staff of the site didn’t immediately discount the possibility. The chap himself referred to it dismissively just once, then completely ignored it, which if anything made it look MORE like he was a non-conscious machine.

    It was all very funny, and as I say it was well over 20 years ago. I didn’t think I’d live to see it literally come true.

    Oh… and the chap’s name – his REAL NAME that he posted under, that he used elsewhere such as on his Youtube channel, years later?

    Justin.

    https://h2g2.com/user/U1711705/conversation/view/F2209694/T699212/page/1/

  2. Alan G. Humphrey says

    The handlers of Justin the AI are going to have to do a “rethink” of how the Catholic Church’s deep and nuanced understanding per Justin’s first response can explain how Galatians 3:28, ” there is no male and female”, is consistent with the current Catholic Church’s doctrine of male only priests.

  3. SeanM says

    The backpedaling Catholic reminds me of a scene from the fantasy novel “Cugel’s Saga” by Jack Vance. A wizard has just given the protagonist Cugel some gifts, and is about to try to convince him to show his appreciation by surrendering a powerful magic scale he carries.

    “I was trained in the old tradition. We found our strength in the basic verities, to which you as a patrician must surely subscribe. Am I right in this?”

    “Absolutely, and in all respects!” declared Cugel. “Recognizing, of course, that these fundamental verities vary from region to region, and even from person to person.”

  4. says

    Looks like this Justin-bot is trying to present itself as The Second Most Interesting Man in the World.

    “I don’t always spout apologetic bullshit for the Catholic Church…but when I do, I only get it directly from the Vatican Archives in Rome. Stay obedient, my friends!”

Leave a Reply

Your email address will not be published. Required fields are marked *