An Architect, An A.I. and a Priest Walk Into a Bar . . .
A Surprising Take on A.I. from the Vatican
In this Post:
What the Vatican Really Thinks About A.I.
Eric Tests the Limits of God's Patience
Why Human Beings Have to be 'The Decider'
Human Flourishing as the Ultimate Motive for All Labor
Well, we're definitely in the end times, folks - I find myself in broad agreement with the Vatican, at least as to their positions on A.I. Yeah, I'm as surprised as anyone. I can't remember ever agreeing with the Vatican on anything, ever. But I decided to read the Vatican's recent epistle: ANTIQUA ET NOVA – Note on the Relationship Between Artificial Intelligence and Human Intelligence.
I haven't read any religious texts in a while. But I used to read them all the time. It was just after I had visited the Hagia Sophia for the first time, and I found myself wondering what it was about religion, that inspired such great architecture. At times when human beings were hardly capable of building any buildings, the ones the built for their gods were massive, beautiful, and permanent. So I just started reading (and in some cases, rereading) all the sacred texts - the KJB, the Talmud, the Bhagavad Gita, the Confucian Analects . . . I was looking for the commandment that said 'pour all your resources and all your architectural brilliance into this kind of building.' I never found it. And yet, through all civilizations, there's this irresistible compulsion to devote the best of Architecture towards the glorification of one's gods. At least, until recently, when we decided to glorify capital instead. Which is arguably the same thing.
I'm not a believer, but I'm grateful that other people are, and that that belief creates and sustains these great buildings. Because great architecture is sorta like our human memory - it reflects the circumstances and passions of its time, and then carries memories forward for future generations.
In a way, architecture performs the same function as the Church itself, and maybe that's why so many of history's greatest buildings were religious in nature.
Both Architecture and the Church sustain. Architecture preserves the tastes, agendas and stories of one generation and carries it forward to the next. Religion charges itself with maintaining a doctrine, based on some really old texts, and keeping it relevant amid wild changes in technology, geography, politics, etc. Doing so requires becoming a 'living' religion, and using the convictions and attitudes of each era as a lens through which to reinterpret those ancient documents.
So I don't know why I was surprised that the Church is now confronting AI, in the way that Architecture is (or should be). Both institutions have much experience in adapting to great changes, having done it for millenia. So I decided to peruse the Vatican's take on AI, and see if this 2,000 year old institution had any wisdom to share.
What surprised me:
It's technically lucid and grounded. Not sure who actually wrote the text, but whoever did understands AI, at a technical level.
It's very human centered, arguing not for the interests of God, or Jesus, the Pope, or the Church itself, but more broadly for the human.
Having re-read it several times now, I can attest that, however you feel about the Catholic Church, religion, etc., it s a pretty good statement on AI. Or, at least, it accords well with my ideas about AI. I'll let you decide for yourself. For those of you without the time/patience to read the whole thing (and its 215 footnotes), here's the highlights:
What the Vatican Really Thinks About A.I.
The Note's basic take is that AI development should be governed by a concern for the human.
The Note acknowledges AI as an 'epochal change' while trying to establish clear boundaries between human and artificial intelligence. The Vatican emphasizes that while AI is a powerful computational tool, it fundamentally differs from human intelligence, which uniquely encompasses body, soul, rationality, all of which speak to, and are derived from, a relationship with God.
The Note gets into the weeds, too. Pope Francis delves into specific challenges across various sectors including healthcare, education, work, and military applications, and attempts to draw ethical guidelines for AI development and deployment in each of those fields.
At its core, The Note insists that human dignity must be the primary criterion for evaluating AI technologies, with moral responsibility always resting with humans rather than machines. While AI will certainly be a powerful tool for human progress, it must be guided by ethical frameworks that preserve human dignity, promotes the common good, and respects humanity's ultimate purpose in relationship with God and others. According to Church doctrine, all human technological progress can be viewed as part of God's will, and part of God's creation, because God imbued humans with the ability to 'tend the garden.' According to Genesis 2:15,
"the Lord God took man and put him in the garden of Eden to tend and keep it."
I'm not sure God had A.I. in mind when writing that phrase, but the Church has taken this to mean that humankind's creations are, broadly speaking, what God intended. As long as we're helping take care of God's creation, then we're on the right side of judgment.
The Note further warns against turning AI into a perceived 'Other' superior to humans in some way – at best, that's idolatry. At worst, because AI is just trained on human output, humans would just end up worshiping themselves
In previous epistle, Pope Francis sums it up well:
In the end, the way we use it [A.I.] to include the least of our brothers and sisters, the vulnerable and those most in need, will be the true measure of our humanity.1
I couldn't agree more.
Eric Tests the Limits of God's Patience
I actually had to read this several times. I was just flabbergasted to find myself in agreement with the Vatican, but hey, here we are. Upon subsequent readings, I realized that I was actually harboring some objections - not in the Vatican's ultimate conclusions, but more in the way that they arrived at those conclusions. That's easily explained by our different belief systems - the Church, being Catholic, and myself, being either agnostic or atheist, depending on how you find me.
Why does that matter? I think because in order to properly regulate the development of AI, we have to assemble the strongest possible arguments with the widest possible appeal. AI is a civilizational issue, and concerns every human on Earth. Our arguments for, against, and around AI have to be, in a sense, doctrine-agnostic, and appeal equally to Catholics, Christians, Jews, Muslims, Hindus and every other faith. I therefore sat down to pick apart the Vatican's arguments - not because I disagreed with their conclusions, but because I actually wanted them to be stronger.
Will talking back to the Vatican earn me heavenly retribution? I dunno, but if I get hit by lightning in the near future, at least you all will know why.
Here's my issues:
The Vatican Uses Linguistics to Circumvent Tough Questions:
That's not really a surprise - a lot of people do this nowadays, at least where AI is concerned. Its a simple trick: you use terms that necessarily, and exclusively, apply to human beings, and then claim that AI can't do or be those things. Well, duh. I'm not sure what that proves. For instance, in arguing that humans have a higher level of intelligence, the Vatican writes:
"Since AI lacks the richness of corporeality, relationality, and the openness of the human heart to truth and goodness, its capacities--though seemingly limitless--are incomparable with the human ability to grasp reality. So much can be learned from an illness, an embrace of reconciliation, and even a simple sunset;" (PP 33)
It is true that much wisdom can be gained from an illness. And machines cannot get sick, they can only break. But one could also gain a lot of wisdom from reading every book that's ever been written, and no human can do that. How does one measure one form of knowledge/wisdom/learning against another? I don't rightly have the answer to that. But one doesn't necessarily strike me as superior to the other. The Church, on the other hand, is definitive. The superior forms of learning are the ones that only humans can do. If that's taken as true, then, yes, only humans would be capable of higher learning.
The Vatican Idealizes Humanity
Granted, I do that too. I do the work that I do because I have a fundamental belief in the goodness of people. But people can be assholes, too. It's convenient to argue that humanity, at its best, is better than AI at X, while ignoring the same comparison when humanity is at its worst. For instance, when discussing AI's incapacity for empathy, the Church writes:
"Unlike the realm of analytical judgment in which AI excels, true empathy belongs to the relational sphere. It involves intuiting and apprehending the lived experiences of another while maintaining the distinction between self and other. While AI can simulate empathetic responses, it cannot replicate the eminently personal and relational nature of authentic empathy."
How often do we experience the kind of empathy that the church describes? For us lucky ones, we can turn to family or intimate friends and find that kind of empathy when we need it. But many have to live their lives entirely without that experience. So at its best, I would agree that human empathy will always trump machine empathy. But in a world so devoid of empathy, shouldn't we consider machine empathy as an alternative? If you had the choice between a human being that was capable of empathy, but didn't show it, and a machine that was incapable of empathy, but could fake it really well, with whom would you prefer to spend your time?
The Vatican has similar takes when it comes to human intelligence, and human moral judgment. I still believe that humans, at their best, exceed whatever machines are capable of. But we're not always at our best, are we?
The Vatican's Arguments Presuppose a Creator.
I'm not sure they did this consciously, but many of their arguments indirectly depend on one's having a belief in a Creator. For instance, the Church spends a great deal of time arguing that humans' perception of 'reality' is superior to that of machines, because machines can only be informed by 'data' while humans are informed by actual 'experiences' with 'reality.' I could pick that apart in several ways, but here I'll just take issue with the idea of 'reality.' What is 'reality'? That's stumped philosophers and scientists and potheads for thousands of years. The only people who ever seem to have confidence that their reality is the reality are A) morons and B) religious people (NB: I am in no way equating the two). Morons feel confidence in their reality because they can't conceive of anything else. Believers feel confident in their reality because in their worldview, a single, authentic reality must exist. If you believe in a Creator, then you must believe that the one, authentic reality is the one created by, and governed by, that Creator. Even if you, personally, don't feel you have a good grasp on reality, you know that someone does (the 'someone' being God, Yahweh, Allah, or other).
Arguing whether humans or machines have the better take on reality is an interesting rabbit hole. Yes, machines don't experience the world in the way that humans do. But they also don't have the cognitive biases that plague humanity. And, they're capable of understanding the physical world on levels that we are not - for instance, being able to see in multiple spectra of light. Fundamentally, though, that argument will go around in circles because no one (except for a deity (that you believe in)) is capable of putting their stamp on one particular version of reality and saying 'This is the correct one.'
I'm not sure it'd be honest to describe the aforementioned as rhetorical flaws. I mean, the Vatican's positions are well-reasoned, as long as that reasoning rests on a foundational idea of a Divine Creator. The Pope believes in one, and I don't.
But even across that chasm of disagreement, the Vatican and I arrive at similar conclusions on two things: the relative moral burdens of AI & humanity, and the question of labor:
Why Human Beings Have to be 'The Decider'
The Note argues that humans must always be in the role 'decider' - humans must make the decisions because only humans are capable of moral judgment, because only humans have a soul, and therefore a relationship with God.
I can't support such a metaphysical rationale. But I do believe that it is extremely unlikely that machines will ever develop a morality. My reasoning is simply that humans are the only form of life with an awareness of our own mortality. We know that we will die some day, and that all of this will end. All of our ambitions, our work, our experiences, will someday just stop. That makes what we do here, now, meaningful. We must act in a certain way, while here on Earth, because we won't always be here.
Machines have no such experience of existence. Even if an AI were to achieve self-awareness, it would certainly know that it cannot 'die.' It can only be turned off, and that's not the same thing, principally because it can always be turned back on again. An AI could, theoretically, live forever.
Paradise Lost explores, at great length, the idea that humans are morally superior to angels, because humans are both mortal and fallible, neither of which applies to angels. Humans face a daily choice of what to do with their limited time, and that gives rise to reflection, judgment and wisdom. Angels (and AI, for that matter) can never reach such a deliberation. Across an infinity of time, the question of whether to do the 'bad' thing or the 'good' thing is meaningless.
Human Flourishing as the Ultimate Motive for All Labor
My most surprising point of agreement with The Note occurred around the topics of creation, labor, utility, and humanism - all topics which have informed my work through all of its permutations. In 2010, I wrote a book called Down Detour Road, with a simple premise: Architecture's cataclysmic fall during the Great Recession (relative to other professional disciplines) was a self-inflicted wound. During the Deconstructivist period, we had excised the 'human' from architecture, and made design about shapes & ostentation instead.
Architecture stopped caring about people, so people stopped caring about architecture.
The solution to Architecture's economic woes, therefore, was to make architecture, once again, about the human - to make architecture that cared about everyday people, their trials, their homes, their families, and the planet on which we all live.
The Note takes a similar stance on A.I. And it is frankly a better one than what I hear from a lot of architectural technologists. I often find myself at odds with those folks, principally because I don't support the idea of using AI to further the current model of architectural practice.
Granted, I can understand the temptation of a working architect to use A.I. to find ways to make his or her day a little easier. But finding a tool which makes doing the wrong thing faster, or easier, or cheaper is often a harbinger of doom. The 'Let's use AI to do what we currently do, just faster and cheaper' mindset has no endpoint, other than commoditization and unemployment. And it doesn't necessarily make architecture better for the humanity we're supposed to be serving.
In the Church's view:
it is essential "that we look for solutions not only in technology but in a change of humanity."[175]A complete and authentic understanding of creation recognizes that the value of all created things cannot be reduced to their mere utility. Therefore, a fully human approach to the stewardship of the earth rejects the distorted anthropocentrism of the technocratic paradigm, which seeks to "extract everything possible" from the world,[176]and rejects the "myth of progress," which assumes that "ecological problems will solve themselves simply with the application of new technology and without any need for ethical considerations or deep change."[177]Such a mindset must give way to a more holistic approach that respects the order of creation and promotes the integral good of the human person while safeguarding our common home
Architecture, too, should reject the 'myth of progress' and the philosophies of extraction. It should center the integral good of the human person, and has a charge to do so for all people, not just the clients.
To the degree that it does, A.I. will be a powerful tool in furthering that mission. In instances where it doesn't, A.I. merely becomes a tool to do the wrong thing faster.
The Note prompts us to see this current A.I. revolution as an opportunity for re-evaluation and pivoting. Rather than asking 'how will A.I. help me with what I'm doing?' or 'will A.I. replace what I'm doing?', ask instead 'Am I currently doing the right thing?' or 'Should I be doing something different?'
That's kinda what I've been saying to architects all this time. There's an opportunity to use this moment to redefine architecture entirely, into the profession that we know it can be.