Question for our hosts - AI generated post policy

BStrauss3 · December 31, 2025 6:15AM

We are seeing more and more posters just dropping AI-generated content (politely slop) into posts without doing any fact-checking or applying critical thinking.

Have PCGS considered a policy statement regarding this?

lermish · December 31, 2025 6:21AM

Let me _definitively _answer this question for you.

PCGS has given this **ZERO **thought whatsoever.

Toss it on the pile of things PCGS should pay attention to and will absolutely ignore because it doesn't affect the top or bottom line positively.

P0CKETCHANGE · December 31, 2025 6:31AM

I asked AI to react to OP's query, and...

Oh, the horror—AI slop invading the sacred grounds of numismatic wisdom! Next thing you know, it'll be hallucinating VAM varieties that don't exist or insisting your cleaned Morgan is "problem-free."

But hey, if we're banning unchecked AI dumps, maybe we should extend it to those "expert" opinions pulled straight from eBay listings or that one blurry Photograde scan from 1987. Fair's fair.

In all seriousness, it's a valid point. Fact-checking is the collector's creed—whether the brain behind the post is carbon or silicon. PCGS mods, your move.

jmlanzaf · December 31, 2025 6:49AM

@P0CKETCHANGE said:
I asked AI to react to OP's query, and...

Oh, the horror—AI slop invading the sacred grounds of numismatic wisdom! Next thing you know, it'll be hallucinating VAM varieties that don't exist or insisting your cleaned Morgan is "problem-free."

But hey, if we're banning unchecked AI dumps, maybe we should extend it to those "expert" opinions pulled straight from eBay listings or that one blurry Photograde scan from 1987. Fair's fair.

In all seriousness, it's a valid point. Fact-checking is the collector's creed—whether the brain behind the post is carbon or silicon. PCGS mods, your move.

But that is the problem, isn't it? There is no difference between AI and human "slop" and no difference between AI and human "wisdom". Who is going to read every post and fact check?

To @lermish point: it's a free forum. It doesn't add to the bottom line of the company. I don't see any reason why they would let it subtract from the bottom line.

lermish · December 31, 2025 6:55AM

@jmlanzaf said:
To @lermish point: it's a free forum. It doesn't add to the bottom line of the company. I don't see any reason why they would let it subtract from the bottom line.

Of course they wouldn't. I'm surprised they have even let the forum continue given the hour or two a week it probably costs them to pay the mods and the thousands of dollars a year in software and server costs.

CaptHenway · December 31, 2025 6:58AM

Was this inspired by the donkey token thread, where AI did leap tall logic with a single bound?

I did hazard an educated guess, labeled as such, based in part on years of junk box diving. It was amazing what you could find in them back in the 60’s and 70’s!

Peasantry · December 31, 2025 7:09AM

"Mom! Mooom!!" 🥲

The_Dinosaur_Man · December 31, 2025 7:18AM

I don't use AI, but I had a friend who does ask the question "which is a better resource, AI or a highly intelligent human?" The results were pretty laughable.

jesbroken · December 31, 2025 7:37AM

I used to enjoy the video 'shorts' on youtube, but now so much and soon maybe all, are AI that, what's the point. When a small child says or does something precious it is worth viewing, but when you have 45 seconds of constant AI cuteness, what's the point. I for one hate the AI trend across the board. JMO
Jim

WACoinGuy · December 31, 2025 8:13AM

Just a little though experiment - judge for yourself whether this feels appropriate or not.

Prompt to AI:

Read this thread (link to this post). Write a response noting a) fantasy / cute photo posts in a serious forum can be an issue b) fact-checking is essential regardless of authorship c) strengths you feel AI can bring to a post

AI response:

I think there are a few different issues getting mixed together in this discussion, and it’s worth separating them so we can talk about the real impact on the forum.

a) Fantasy / “cute” AI photo posts in a technical forum
The U.S. Coin Forum has always worked best when posts stay anchored in numismatics. Lighthearted content has its place, but when AI‑generated fantasy images or “cute” photo dumps start showing up in threads where people are trying to identify varieties, evaluate surfaces, or discuss history, it can dilute the signal‑to‑noise ratio. The problem isn’t AI per se—it’s context. A serious attribution thread isn’t the right place for synthetic imagery that looks like a coin but isn’t one.

b) Fact‑checking matters no matter who (or what) wrote the post
Whether a post is written by a human, an AI model, or a parrot trained on Red Book tables, the responsibility is the same: verify before posting. AI can be wrong with confidence, but so can humans. We’ve all seen incorrect “expert” answers, misremembered mintage figures, and eBay‑sourced mythology. The tool isn’t the issue—unvetted information is. If someone uses AI to draft a post, the burden is still on the poster to check dates, varieties, diagnostics, and historical claims before hitting “submit.”

c) What AI can bring to the table—when used responsibly
AI isn’t a replacement for expertise, but it can be a useful assistant if the user applies judgment. For example:

Summarizing long historical sources so a thread can get to the relevant facts faster.
Drafting clearer explanations for beginners who are still learning terminology.
Providing structured comparisons (e.g., differences between two die marriages) that a human can then refine.
Helping organize information—timelines, bullet points, references—so discussions stay focused.
Spotting inconsistencies in a claim or dataset that a human can then investigate.
Used thoughtfully, AI can help raise the clarity and accessibility of posts. Used lazily, it can add noise. The difference isn’t the technology—it’s the discipline of the person using it.
In short: context matters, verification matters, and good tools used well can strengthen the forum rather than weaken it."

jacrispies · December 31, 2025 8:42AM

I believe developing a culture of specifically noting "the following is AI generated" is the best option. AI isn't inherently bad, for it has its use. But in the hands of the lazy, it can be aimless and misleading. Reading a post that begins with "the following is AI generated" would be very helpful, and we can make judgements from there.

jmlanzaf · December 31, 2025 9:09AM

@lermish said:

@jmlanzaf said:
To @lermish point: it's a free forum. It doesn't add to the bottom line of the company. I don't see any reason why they would let it subtract from the bottom line.

Of course they wouldn't. I'm surprised they have even let the forum continue given the hour or two a week it probably costs them to pay the mods and the thousands of dollars a year in software and server costs.

Yes. I was agreeing with you

jmlanzaf · December 31, 2025 9:11AM

@jacrispies said:
I believe developing a culture of specifically noting "the following is AI generated" is the best option. AI isn't inherently bad, for it has its use. But in the hands of the lazy, it can be aimless and misleading. Reading a post that begins with "the following is AI generated" would be very helpful, and we can make judgements from there.

But that should be true of EVERY post. How many actual citations do we get?

pcgs_education · December 31, 2025 9:18AM

Hello everyone. We understand that the copying and pasting of AI-generated responses can be frustrating. However, this is a public forum, and the use of AI tools does not violate our posted guidelines. As such, we will not be monitoring individual posts for AI usage.

We strongly encourage best practices, including fact-checking information before posting and before accepting any content shared by others as factual.

jacrispies · December 31, 2025 9:42AM

@jmlanzaf said:

@jacrispies said:
I believe developing a culture of specifically noting "the following is AI generated" is the best option. AI isn't inherently bad, for it has its use. But in the hands of the lazy, it can be aimless and misleading. Reading a post that begins with "the following is AI generated" would be very helpful, and we can make judgements from there.

But that should be true of EVERY post. How many actual citations do we get?

Not always. Any information in the hands of the lazy can be aimless and misleading. But, many sources are "reliable enough" for the use of this forum. Many of us in the coin industry work from our memory and don't necessarily memorize every source/catalog/presentation we run into throughout our lives. In many cases that is usually good enough, keeping that it bears witness to the experience and observation of others.

For the case of AI, unchecked is less reputable than wikipedia in my opinion. What do you think?

Morgan White · December 31, 2025 10:21AM

As long as they don't ban memes, I'm good.

Morgan White · December 31, 2025 10:24AM

@jmlanzaf said:

@lermish said:

@jmlanzaf said:
To @lermish point: it's a free forum. It doesn't add to the bottom line of the company. I don't see any reason why they would let it subtract from the bottom line.

Of course they wouldn't. I'm surprised they have even let the forum continue given the hour or two a week it probably costs them to pay the mods and the thousands of dollars a year in software and server costs.

Yes. I was agreeing with you

jmlanzaf · December 31, 2025 10:30AM

@jacrispies said:

@jmlanzaf said:

@jacrispies said:
I believe developing a culture of specifically noting "the following is AI generated" is the best option. AI isn't inherently bad, for it has its use. But in the hands of the lazy, it can be aimless and misleading. Reading a post that begins with "the following is AI generated" would be very helpful, and we can make judgements from there.

But that should be true of EVERY post. How many actual citations do we get?

Not always. Any information in the hands of the lazy can be aimless and misleading. But, many sources are "reliable enough" for the use of this forum. Many of us in the coin industry work from our memory and don't necessarily memorize every source/catalog/presentation we run into throughout our lives. In many cases that is usually good enough, keeping that it bears witness to the experience and observation of others.

For the case of AI, unchecked is less reputable than wikipedia in my opinion. What do you think?

Wikipedia? Yes. Other random internet sources... not so much.

JBK · December 31, 2025 10:33AM

I don't even like Wikipedia sometimes. 🤨

CaptHenway · December 31, 2025 12:32PM

@jacrispies said:
I believe developing a culture of specifically noting "the following is AI generated" is the best option. AI isn't inherently bad, for it has its use. But in the hands of the lazy, it can be aimless and misleading. Reading a post that begins with "the following is AI generated" would be very helpful, and we can make judgements from there.

Excellent idea.

Question for our hosts - AI generated post policy

Comments

Leave a Comment