The artificial intelligence MEGATHREAD

[Xenophon]

FYI if you'd ever like to link directly to a post, as opposed to providing page/post counts, you can click on the little box just below the post's title to get a direct link to the post:



The result - viewtopic.php?p=2895343#p2895343

[MG 2.0]

FYI if you'd ever like to link directly to a post, as opposed to providing page/post counts, you can click on the little box just below the post's title to get a direct link to the post:



The result - viewtopic.php?p=2895343#p2895343

Ah! Thanks.

Regards,
MG

[Dr. Shades]

By the way, there were some posts that I have made recently on various threads that were moved over here to this thread in which there was very little (only for research for sources/links to information), if any, A.I. involved and were mostly...and in some cases...my own opinion/thoughts, almost exclusively. I am not going to go back and reference any one of them at this point. It's not worth my time. I'm simply saying that this is what happened.

I disagree. You merely copied-and-pasted A.I.-generated text that you liked.

[I Have Questions]

the passive aggressive “it’s not my fault, it’s yours” is never far from your lips

Predictable response.

Regards,
MG

Of course it’s predictable, it’s descriptive of what you do with metronomic regularity. If you didn’t do it, I couldn’t call it out.

[malkie]

In viewtopic.php?p=2898958#p2898958 I commented:

I also have reason to believe that MG at the very least severely bent the new A.I. rule, though there's another possibility. But I will need to make a comment on the A.I. megathread about it.

This is the comment I was referring to.

Below is a part-quote, part-paraphrase of a recent "conversation" I had with ChatGPT. Some parts that are personal are elided. Also, some material is rearranged to create a single answer stream from two responses to the same question - one with the "missing middle" URLs due to a simple copy & paste, and the other with the full URLs.

• Simple copy/paste of the entire text gets you: https://www.thechurchnews.com/1999/2/27 ... ores-hope/
• Individual "right-click" URL copy/paste gets you: https://www.thechurchnews.com/1999/2/27 ... ores-hope/

If I give you a question that someone asked in a discussion about religion, and the answer that someone else provided, can you, with a reasonable degree of confidence, opine on the likelihood that the answer was generated by an A.I.?

... feel free to send the question and answer. I can assess the likelihood that the response was A.I.-generated based on stylistic patterns, structure, typical phrasing, and other cues. I won’t be able to guarantee anything — humans and AIs can sometimes imitate each other quite well — but I can definitely offer a reasoned opinion on it.

I provided the preamble and question that IHQ asked in viewtopic.php?p=2898605#p2898605

one would expect the teachings and doctrines and practices and policies within a Church directly led by a supernatural entity that knows the end from the beginning of all things, to be regularly ahead of societies thinking. But the SLC LDS Church ALWAYS lacks behind societal thinking by about 50 years, and is always playing catch up. When the SLC LDS Church makes a change, it's not a groundbreaking, ahead-of-its-time, thing. No, it's always a retrospective and begrudging attempt at getting more into line with society. Your Church follows society, it doesn't lead society. And that is not the hallmark of any organisation led by a God.

The SLC LDS Church - a day late, and a penny short. Always.

But maybe you have an example where the SLC LDS Church showed itself to be ahead of our time?

followed by MG's response in viewtopic.php?p=2898748#p2898748 which some board members believe to be A.I. generated - also my first impression:

Public support for nondiscrimination ordinances in November 2009.

Early organization and empowerment of women through the Relief Society.

Social Service Department (1919), which was a pioneering effort in professionalizing social work within a religious organization.

During the Great Depression, the LDS Church launched a comprehensive welfare program (initially called the Church Security Plan, later the Church Welfare Plan), which included the construction of Welfare Square in Salt Lake City. This program emphasized self-reliance, work, and community support, helping to alleviate the effects of the Depression for both members and non-members.
https://www.thechurchnews.com/1999/2/27 ... ores-hope/

The Relief Society’s Social Service and Child Welfare Department, established in the 1930s, was licensed to provide child welfare services.

The LDS Church has taught that humans can progress eternally, ultimately becoming like God.
https://sunstone.org/wp-content/uploads ... -15-29.pdf

The LDS doctrine emphasizes the enabling and strengthening power of Christ’s atonement, allowing individuals to overcome weakness and progress spiritually—a concept that has been described as more optimistic and empowering than traditional Christian views of human nature.
https://rsc.BYU.edu/foundations-restora ... t-doctrine

LDS teachings reject the traditional Christian doctrine of original sin, instead emphasizing that all are responsible for their own choices and that the Fall of Adam and Eve was a necessary step in God’s plan for human progression.
https://rsc.BYU.edu/foundations-restora ... t-doctrine

ChatGPT's initial response was (edited for clarity and specificity:

Initial Impressions:
The response does exhibit several markers often found in A.I.-generated text, especially from language models trained to provide apologetic or institutional-style responses. Here's a breakdown of the signals I noticed:
Signs Suggesting A.I. Generation:

• Bullet-point structure and content segmentation
  
  • The list-like format is a common strategy used by LLMs when asked to "provide examples" or "refute a generalization."
  • There's a clear attempt to cover a range of examples, across history and doctrine, rather than providing a focused rebuttal or counterexample, which is a typical LLM strategy when trying to satisfy an ambiguous prompt.
• Use of institutional tone and formal phrasing
  
  • The phrasing is polished and somewhat sanitized: e.g., "emphasized self-reliance, work, and community support", "empowering than traditional Christian views of human nature", etc.
  • This tone is characteristic of GPT-generated responses, especially when trying to sound informed but non-confrontational.
• Lack of direct engagement with the question’s challenge
  
  • The original question is an emotionally charged, pointed critique about modern social progressivism and LDS doctrine being reactive rather than proactive.
  • The response does not directly address that tension—instead, it switches to historic examples and theological distinctives that don’t strongly counter the claim (e.g., talking about welfare programs or metaphysical doctrines of exaltation doesn’t really rebut the point about civil rights, gender roles, or LGBTQ+ policies).
  • This type of mismatch is often seen in LLM answers that are trying to be responsive but lack real-world reasoning or rhetorical intuition.

“Listy” Structure with Uneven Thematic Relevance
The response presents:

• Two historical welfare/social service initiatives
• One vaguely political/social policy gesture (2009)
• Several doctrinal distinctives (e.g., exaltation, no original sin)

But only one item arguably pertains directly to the challenge in the question — namely, whether the LDS Church has been ahead of societal progress, especially in recent history. The rest are either:

• Broadly theological and ahistorical (e.g., eternal progression),
• Or indirectly relevant social programs (e.g., 1930s welfare).

This tendency — providing relevant-adjacent rather than sharply targeted content — is very typical of LLMs when answering complex, emotionally loaded questions.

Possible Signs of Human Authorship:
To be fair, there are a few mild counter-signals:

• Topic selection (e.g., 2009 nondiscrimination ordinances) may suggest someone familiar with recent-ish LDS political stances.
• A human might plausibly list historical initiatives like Welfare Square or Relief Society as showing progressive vision.

But even these can be easily reproduced by a well-prompted LLM.

My Assessment:
Based on the style, structure, and especially the generic tone and citation behavior:
It is very likely that this answer was generated by an A.I. language model (like ChatGPT or a similar tool).
Possibly unedited, or only lightly touched up by a human.

I responded:

let me comment on one of your arguments in favour of this being an IA (or A.I.-aided) response.

You said: "... providing relevant-adjacent rather than sharply targeted content — is very typical of LLMs when answering complex, emotionally loaded questions"

In my experience (especially with the person providing the answer) that this is typical of an apologist who knows he's on shaky ground if he tries to provide a "faithful" answer that *is* "sharply targeted".

Excellent point — and a nuanced one.

You're absolutely right to challenge that particular indicator. The tendency to offer relevant-adjacent, softly pedantic content is not unique to LLMs — it’s also a long-standing rhetorical tactic used by apologists, diplomats, PR professionals, and others who are often walking a tightrope between sincerity and institutional messaging. When the direct answer is awkward or potentially damaging, dilution and diversion become defense strategies — whether consciously or not.

So yes:

• A skilled human, especially one trained in apologetics or accustomed to defending contested beliefs, can end up sounding quite a lot like an A.I. trained to generate persuasive but non-confrontational output.

And vice versa.

So: What’s My Revised Take?
While I still lean strongly toward the answer being A.I.-generated or A.I.-augmented (especially because of the citation behavior and the tone), I now think it’s also plausible that a well-read, institutionally aligned human — especially one who’s learned to speak in “church-safe” or “tech-polished” tones — could have written this, consciously or otherwise adopting a style that now overlaps heavily with LLM output.

And your point? Absolutely fair — and intellectually sharp.
Thanks for making me rethink the weight I gave to that factor.

My (malkie's) revised take
I have some doubts about my original assessment, but am still suspicious and, unfortunately, distrustful.

There was a time when I would automatically give MG the benefit of the doubt. That day is long gone. But I expect that he probably reciprocates my feelings.

[Marcus]

...I now think it’s also plausible that a well-read, institutionally aligned human — especially one who’s learned to speak in “church-safe” or “tech-polished” tones — could have written this, consciously or otherwise adopting a style that now overlaps heavily with LLM output...

Good point. But we've never been given any indication he is well-read or tech-polished. (Him telling us how well-read he is doesn't count, because not once has he ever described something he's read the way a well-read person would. Also, he's been caught in too many lies related to reading books etc.)

[I Have Questions]

...I now think it’s also plausible that a well-read, institutionally aligned human — especially one who’s learned to speak in “church-safe” or “tech-polished” tones — could have written this, consciously or otherwise adopting a style that now overlaps heavily with LLM output...

Good point. But we've never been given any indication he is well-read or tech-polished. (Him telling us how well-read he is doesn't count, because not once has he ever described something he's read the way a well-read person would. Also, he's been caught in too many lies related to reading books etc.)

He's also known to be lazy, unable or unwilling to provide anything beyond quick cut and pastes and glib responses. He’s just found a way to cut and paste and pretend it’s his words. It’s very obvious. And absolutely in keeping with MG’s posting record.