[mbork_pl]

Looks like the article is behind a paywall, so I only read the abstract, but it looks that there is a significant flaw in this study. There are really many really various religions out there, and the results _might_ depend heavily on the choice of religion. As a thought experiment, consider a fictional example religion where A.I. is treated as a god, compared to the religions of the Dune universe (where A.I. is a taboo).

[debok]

I also only read the abstract.

The argument seems to be: Belief in God implies a belief in the fallibility of humans, which leads to a reduced reliance on humans and therefore increased willingness to accept AI recommendations.

Most of the logic in that argument checks out. I just don't understand the last step in the logic. How do decreased reliance on humans lead to increased willingness to accept AI?

I'm a religious person myself, and my argument would be to not trust AI all that much. It is a creation by fallible human beings trained on fallible human data.

[mbork_pl]

> I'm a religious person myself, and my argument would be to not trust AI all that much. It is a creation by fallible human beings trained on fallible human data.

Good point. How is A.I. different (in this respect, and possibly in some other ones, too) from a really huge, partially pseudo-random spreadsheet?

We need to remember, though, that we live in a relatively small bubble where A.I. is perceived as (more or less) what it is. For a layman, both an A.I. and a huge spreadsheet may be similarly beyond comprehension.

I will always remember talking to a very nice old lady (it could have been about 25 years ago), who told me that computers can read people's minds. She saw someone type a few characters on a computer and suddenly a whole page appeared. Since she had a mental model of a typewriter (which has no memory), the logical conclusion of text appearing without a person physically typing it was that the machine read his mind. In fact, that made perfect sense!

We often have no idea how weird (from our POV) the people's mental models of computers are.

[debok]

Yeah, you make a good point. It reminds me of that Arthur C. Clarke quote: "Any sufficiently advanced technology is indistinguishable from magic." I guess if you see AI as something magical, it could easily start occupying a religious or God-like space in your psyche.

[orhmeh09]

You don't need to read the article, just search online a little. It probably took you longer to write that post for me to find this: https://osf.io/fdh4m

They assess belief in god in different ways. Details are found within the 8 preregistered study PDF documents.

[ben_w]

Different religions is potentially interesting; though now I'm also wondering if this is cultural "God salience" or serious "God salience" (the way the British do Easter and Christmas even though mostly not taking it seriously).

-

"""Studies 3 and 4 demonstrate that the reduced reliance on humans is driven by a heightened feeling of smallness when God is salient"""

I'd expect this to be less true for pantheons.

"""followed by a recognition of human fallibility."""

In pantheons, I'd expect this to vary by the nature of the god/goddess in question; Greek revivalism probably has different answers when considering Athena vs. Dionysus.

"""Study 5 addresses the similarity in mysteriousness between God and AI as an alternative, but unsupported, explanation."""

Yeah, that feels plausible.

"X and Y are mysterious, perhaps they're the same?" seems common for any {X, Y} — AI, consciousness, quantum mechanics, god, evolution, prime numbers, art, …