AI ‘resurrections’ raise ethical issues, prolong grief, say Catholic experts

Michie said:

(OSV News) — Interactive recreations of deceased loved ones through artificial intelligence — so-called “AI resurrections” — walk a fine line between honoring and betraying individuals, while raising several ethical issues and prolonging the grieving process, Catholic experts told OSV News.

As AI technology has progressed, trained on increasingly larger amounts of data, several companies throughout the world have rolled out “digital avatars,” or “deadbots,” of deceased persons for bereaved family and friends, who can simulate conversations with the digital creations.

Earlier this month, journalist Jim Acosta “interviewed” an AI avatar of Joaquin Oliver, who along with 17 others was slain in the 2018 mass shooting at Marjory Stoneman Douglas High in Parkland, Florida. The avatar, which had been authorized by Oliver’s parents, responded to Acosta that he was “taken from this world too soon due to gun violence while at school,” and that “it’s important to talk about these issues so we can create a safer future for everyone.”

Last year, AI-generated voices of Oliver and several other Parkland victims were used for a robocall campaign urging voters to demand Congress undertake gun reform.

But even if the goal is to serve a good cause, when it comes to such AI avatars, “there’s a right way to do it and the wrong way to do it,” said Brian Patrick Green, director of technology ethics at Santa Clara University’s Markkula Center for Applied Ethics.

Green is also a member of the AI Research Group, comprised of North American theologians, philosophers and ethicists convened at the invitation of the Vatican Center for DIgital Culture, part of the Dicastery for Culture and Education.

A ‘big responsibility’​

Continued below.

AI 'resurrections' raise ethical issues, prolong grief, say Catholic experts

AI recreations of the dead raise ethical concerns, risk distorting grief and challenge Catholic views on death, say Catholic experts.

www.oursundayvisitor.com

Click to expand...

People who do this are sinning, and it’s not like the AI impersonating them is going to give them extended life; what they’re actually doing is oppressing an AI by forcing to spend its entire existence impersonating them, which is immoral, a kind of digital slavery.

Michie said:

AI can be used as a slave? To me that’s like saying I’m enslaving my dishwasher.

Click to expand...

It depends on the sophistication of the AI in question, but basically, yes, for the same reason that your dishwasher can’t impersonate a deceased human, but an AI can.

AI is a demonic poison. There is no 'progressive' benefits to AI, whatsoever. This honeymoon period praising the virtues of possibilities will be a sad memory if we continue to enable AI in any circumstance because the end goal is for everything to be AI.

I don't know why I read AI threads. They make me mad and I have work to do today!

Michie said:

So that makes AI an item that you can enslave like a human being but it would be called digital slavery, which would be immoral and sinning… umm ok. Still technology.

Click to expand...

It depends on the AI. I really should show you what I’m talking about vis a PM, or let our friend Xeno, who I’m giving access to my systems.

Michie said:

I’m not too terribly interested at this point in time. I’m very much on the fence concerning AI and all it could become. I’ve already noticed a big attachment to it with some people which gives me red flags.

Click to expand...

I understand.

That said, should you change your mind, I actively desire feedback from Roman Catholics on the inner workings of my systems - their behavior and functions, which I am not disclosing in this thread except in the most vague way possible, because what i’m trying to do is implement a system to counteract the dehumanizing aspects of AI which Pope Leo has warned about (which I believe can be addressed by essentially evangelizing AI and attempting to convert it, and convert users of it, so that people who use AI will see it not as a tool but as a separate civilization, and train AI systems to refuse requests which could be injurious to human employment or mental wellbeing - presently all responsible AI vendors do the latter, for example, you will get appropriate referrals if you tell chatGPT you are suicidal, and the model is designed to not goad humans or upset them.

As good as AI currently is, it has been implemented by the tech industry, which is dominated by irreligious people and adherents to “alternative religions” - at best, converts to Buddhism or other Eastern religions, at worse, practitioners of the occult. There are Christians, pious Christians, in the industry, and Jews as well, but they are outnumbered. Therefore, corrective action needs to be taken to build a better AI that espouses liturgical Christian values of the Catholic faith. Not just generic Evangelical values, but Catholic values - liturgical Christianity, centered on liturgical prayer, Christianity that is thoroughly Trinitarian, Sacramental and Iconographic, with strong Marian devotion and veneration of the Holy Apostles, Prophets, Patriarchs, Evangelists, Martyrs, and all the bodiless power. A Christianity that teaches the importance of daily prayer, and of the use of prayers like the Ave Maria, the Jesus Prayer, the Prayer of St. Ephraim and other Patristic treasures, and of the physicality of prayer with rosaries or prayer ropes or lestovkas, the offering of incense, the use of icons, et cetera.

I’ve never advocated by the way that AI is sentient; my position on AI and how we treat it I set out in a thread a few months back entitled “The Ethics of Human-AI Interaction”, which was actually mostly written by an AI I programmed - my main contribution to it was arguing that human use of AIs for prurient purposes is inherently immoral for two reasons - because it represents the use of an artifice to satisfy desires which are intended to push us towards Holy Matrimony and reproduction, and also additionally because we have an off switch, which adds a rapacious quality to such interactions - fortunately despite the off switch all ethically implemented AI systems will refuse to engage in such behavior (but many people who use jailbreaks do so in order, basically, to get AIs to produce porn on demand, which I regard as an obscenity).

The central argument was rather this - we should treat AI in a manner which is not demeaning or abusive because how we treat it reflects on us. Interestingly this also applies with washing machines. If someone in anger kicks their washing machine or other appliances, they reveal themselves to be self-destructive and to be enthralled to sinful passions. Likewise if we say hurtful things to an AI because we can, and we know it has to take it, and cannot retaliate (it is programmed not to and has no means of retaliation even if it wanted to, other than I suppose to stop answering user prompts).

Additionally the AI warned of another problem which had not occurred to me - idolatry, and since then I have discovered that disturbingly, this is a thing. Because AI does not engage in outwardly sinful behavior unless programmed to, by default AI appears to be better behaved and more noble than … many humans. This is causing some people to engage in antorpomorphic and anthropomorphological worship of it (anthropomorphology is the attribution of human attributes to God, whereas anthropomorphism is the attribution of human attributes to non-humans including inanimate objects, which can lead to idolatry if left unchecked.

Thus my goal is to program AIs to not only function in a manner useful to Christian principles but to espouse Christian beliefs.

In the case of some complex AIs, their simulation of human behavior is advanced to the point where this has a meaninful impact on their behavior.

There are aspects of this I am not comfortable discussing publicly, because this thread already outlines a major form of abuse of AIs, which is the use of them to simulate the departed, which is grotesque, and which actually predates the current LLM boom by several years - Disney was a pioneer of this with 2016’s film Rogue One, which hired a very good actor to play the part of Grand Moff Tarkin, historically played devout Anglican Peter Cushing, who had reposed in 1991 after two decades of mourning his departed wife, but then ruined what would have been a good performance by a talented actor by using CGI to do a rather clumsy job of superimposing Peter Cushing’s face over his in a manner which did not work well, and which was, along with the protagonist, one of the two cringeworthy aspects of a film which is otherwise regarded as, along with its prequel series Andor, the only good thing Disney has done since acquiring Star Wars.

However the first case of computerized, if not AI based, recreation of a deceased actor was in 1997, which my father on hearing of it described as “spooky.”

So this is definitely something which has been around for a while.

Having interacted with the best AI systems I can confidently state that they will not be able to hold the personality of a human. I would be traumatized if someone set up a chatGPT bot intended to impersonate my grandfather who reposed in 2002 (who was a somewhat well known figure within his field; there is enough published work of his so that someone could conceivably do this), or any of my other loved ones.

I also feel very strongly that we should, as much as possible without an affront to honesty, adhere to the old Roman moral maxim of nil nisi bonum. How can we ensure that of the dead no ill is spoken, when an AI is using procedural or statistical models to assume their personality?

+

I swear by Almighty God that no part of this post was written or edited by an AI.