New forum rule regarding AI content

[Scott5114]

The staff proposes adding a new rule to the forum for calendar year 2026 to regulate the usage of automatically generated ("AI") text. The text of the proposed rule is as follows:

What's not allowed:

Direct copy and pastes of bulk unedited output from or conversations with a generative artificial intelligence (AI) chatbot or other automatic text generator. The purpose of the forum is to facilitate discussion between humans interested in road transportation and related subjects. Long-term storage of machine-generated text is not within the forum's remit. If someone wishes to read machine-generated text, they can use the widely-available tools that exist to generate the text themselves at any time.

The question of whether or not to enact such a rule has been answered—we are going to do it in some form or fashion, owing to the fact that we have observed that the forum membership as a whole has little appetite for consuming AI-generated text content—however, I would like to survey the membership's opinion on the wording of the proposed rule to ensure that it's consistent with community standards, and to sanity-check it to ensure that there are neither unintended loopholes nor unintended banning of content the community would find generally acceptable.

What is still allowed with the wording as proposed:
- Machine-generated images. In my experience, these are mostly used on the forum for memes, visual puns, and other jokes. I see no reason to ban these at this time, but if the usage of generated images changes (especially if they begin to be used in a misleading or deceptive manner) we may revisit this in the future.
- Using a text generator to create a first draft of a response, which the poster then refines into a finished post. When done well, this would not appear to be AI at all, and some writers may find it useful to utilize a text generator as a first step in organizing their own thoughts. This is generally fine since there is a degree of human oversight on the use of the text generator, and the thoughts and opinions of the human user are still the focus of the post.

Mostly what we are trying to avoid are posts along the lines of "Here is what my favorite AI says about this" followed by a swath of computer-generated text.

Barring any major issues being pointed out, which would require a rewrite of the rule and further discussion, it will be enacted on January 1, 2026, at midnight EST.

[LilianaUwU]

Ban it all.

[hotdogPi]

There have been a few AI post dumps that are "look how wrong it is" rather than agreeing with what it says. I think those should continue to be allowed.

[Scott5114]

There have been a few AI post dumps that are "look how wrong it is" rather than agreeing with what it says. I think those should continue to be allowed.

I would be interested in seeing a proposed amendment which would allow this—perhaps a carve-out for commentary on the output? (On the other hand, there are ways to do a "look at how wrong it is"-type post with commentary without copying and pasting the entire output, e.g. selective quoting of the objectionable parts, so perhaps the proposed rule is adequate as-is?)

[formulanone]

There have been a few AI post dumps that are "look how wrong it is" rather than agreeing with what it says. I think those should continue to be allowed.

Could you explain this in clearer detail or give an example?

[gonealookin]

Generally it should be similar to policy regarding quoting a newspaper article.

Nobody should ever copy and paste an entire article.  Copy a few sentences or a single paragraph that go to your point, paste that into a quote box with ellipsis if necessary, then come out of the quote and state in your own words why that point should be of interest to forum members.

[NE2]

But what if I use autocomplete to get a new job for a while now?

[PNWRoadgeek]

But what if I use autocomplete to get a new job for a while now?

Exactly! I think we need to settle down here, AI requires a lot of work! Autocomplete, doesn't.

[Molandfreak]

There have been a few AI post dumps that are "look how wrong it is" rather than agreeing with what it says. I think those should continue to be allowed.

I would be interested in seeing a proposed amendment which would allow this—perhaps a carve-out for commentary on the output? (On the other hand, there are ways to do a "look at how wrong it is"-type post with commentary without copying and pasting the entire output, e.g. selective quoting of the objectionable parts, so perhaps the proposed rule is adequate as-is?)

Does SMF have a code where you can hide text behind a "spoiler" with a brief description of the content? Other forums I have used in the past had this feature, and I suppose it could be a good compromise if the full context is necessary.

[Scott5114]

There have been a few AI post dumps that are "look how wrong it is" rather than agreeing with what it says. I think those should continue to be allowed.

I would be interested in seeing a proposed amendment which would allow this—perhaps a carve-out for commentary on the output? (On the other hand, there are ways to do a "look at how wrong it is"-type post with commentary without copying and pasting the entire output, e.g. selective quoting of the objectionable parts, so perhaps the proposed rule is adequate as-is?)

Does SMF have a code where you can hide text behind a "spoiler" with a brief description of the content? Other forums I have used in the past had this feature, and I suppose it could be a good compromise if the full context is necessary.

Not that I'm aware of (there may be an extension to provide it, though, as this functionality is provided by the base HTML <details> tag now, so it would be fairly simple to provide a BB code with one-to-one mapping to it).

[Scott5114]

But what if I use autocomplete to get a new job for a while now?

Anyone who uses autocomplete will be in the same paragraph as the other one of the most common first names on the way home from the house at the moment but I will be nice to people who are clueless enough to do that but forgetting to actually send the message to you is not a problem for me to get the last of the day after MLK days of work and I will be nice to you and your family and family and family and family and family

[vdeane]

There have been a few AI post dumps that are "look how wrong it is" rather than agreeing with what it says. I think those should continue to be allowed.

I think it depends on what's being included.  That one paragraph of someone asking for directions from Sesame Street to somewhere in DC and getting something that included the Cookie Monster Expressway was hilarious, but I have no appetite for wading through something like the pages of AI text that were posted to a new roundabout thread.

[PColumbus73]

I think it's going to be very in situ determining what kind of AI is allowable. I would imagine users trying to skirt the rule by using the AI generated text and then changing certain words or phrases to sound more authentic.

If we had to boil it down, I think it's going to depend on the subject matter is "Fiction" or "Non-Fiction". If we use the "Sesame Street to DC" example, if the intent is obviously for humor, satire, or parody, or even critical discussions of AI as a tool, then it's possible to permit them, but would be very case-by-case.

However, if we have something like the Francis Scott Key Bridge, where Generative AI is being passed off as factual information, I think that's where AI would need to be outright banned. Because we can't have honest, factual discussions about very non-humorous, serious topics when one user wants to feed replies into a chatbot to produce counter-arguments, or even responses that flatter said user.

Futhermore, posts like this feel like they detract the spirit of the forums. There is often snark and bickering between users here, but getting an AI to generate rebuttals is at the very least poor sportsmanship, and in worse cases bordering on harassment.

[Max Rockatansky]

One particular user is showing exactly what AI is capable in the wrong hands.  I say ban all use of it.

[Beltway]

I am already having discussions with the owner of this forum, and have been going back a few months.

I have been quite critical of the behavior of 8 posters on this forum.
-- Massive numbers of one-liner posts that add nothing any discussion.
-- Massive numbers of garbage posts with images and cartoons
-- A concerted effort by about 6 posters, including government employees, to blow up and shut down any negative discussion about the Baltimore Key Bridge. Note that they engineered the locking of three threads about the bridge.

How about a forum rule against those far greater problems?

-- Now focusing on three MS Copilot queries that I posted and identified as such. TIOLI

Moderators are not professional communicators, the are volunteers that spend a lot of time in that job. Often the only fast  way to slow down a discussion that is out of control -- is to lock the thread -- and those netcop posters know that.

I welcome the participation of the owner Alex in this thread.

[Scott5114]

Alex has had access to the proposed rule for two weeks already during our internal review period. He voiced no objection to it during that time, and I see no reason to expect that to change.

[Molandfreak]

I am already having discussions with the owner of this forum, and have been going back a few months.

I have been quite critical of the behavior of 8 posters on this forum.
-- Massive numbers of one-liner posts that add nothing any discussion.
-- Massive numbers of garbage posts with images and cartoons
-- A concerted effort by about 6 posters, including government employees, to blow up and shut down any negative discussion about the Baltimore Key Bridge. Note that they engineered the locking of three threads about the bridge.

How about a forum rule against those far greater problems?

-- Now focusing on three MS Copilot queries that I posted

Moderators are not professional communicators, the are volunteers that spend a lot of time in that job. Often the only fast  way to slow down a discussion that is out of control -- is to lock the thread -- and those netcop posters know that.

I welcome the participation of the owner Alex in this thread.

[Jim]

I have no concerns about the new rule.

However, I don't see why anyone would want to use AI to generate something for the purposes of posting it on this forum.  I probably only follow 5% of the threads so maybe I'm missing the threads that demonstrate the problem for which this is a proposed solution.

For the OP's example of using it to organize your own thoughts to get a post started, I would hope people choose to organize their own thoughts with their own brains.  The exercise is probably good for it.

[Beltway]

Alex has had access to the proposed rule for two weeks already during our internal review period. He voiced no objection to it during that time, and I see no reason to expect that to change.

Nevertheless, Alex will continue to have my opinions on the conduct of certain individuals on this forum.

[Scott5114]

For the OP's example of using it to organize your own thoughts to get a post started, I would hope people choose to organize their own thoughts with their own brains.  The exercise is probably good for it.

When we were drafting a similar rule at the wiki, an editor there commented that he was not confident in his writing skills, so he would do all of the research to collect sources first, then have the AI generate the text, then correct it to match his research and add the sources in. I can think of no real reason to object to someone doing something like that, even if it's entirely foreign to the way I do things.

[zzcarp]

I use AI regularly in my communications and to organize my thoughts. Often I'll put the gist of what I'm looking to say and ask advice. It can put together some surprisingly good data that you can edit and make your own. I have used it to analyze dating site messages and profiles, and I just used it with an HR-like email from my manager.

That said, you rarely if ever can get content that doesn't need at least a little editing. Perhaps once in a great while. So if there is longform information that is clearly false and clearly AI-generated without editing, I think that is a good idea. I think we just need a carveout for posters editing or organizing their thoughts (which, I think our moderation team would do).

That said, I've never used AI to post on this forum.

[Bruce]

I'm in favor, and would also approve of a rule that doesn't allow for excessive use of AI-generated imagery because we know it could quickly spiral out of control.

[LilianaUwU]

I am already having discussions with the owner of this forum, and have been going back a few months.

I have been quite critical of the behavior of 8 posters on this forum.
-- Massive numbers of one-liner posts that add nothing any discussion.
-- Massive numbers of garbage posts with images and cartoons
-- A concerted effort by about 6 posters, including government employees, to blow up and shut down any negative discussion about the Baltimore Key Bridge. Note that they engineered the locking of three threads about the bridge.

How about a forum rule against those far greater problems?

Beltway, I'll say it in the nicest way possible: if eight people are calling you out, then you should reflect on yourself and think that maybe there is something you should change about yourself.

[Beltway]

I am already having discussions with the owner of this forum, and have been going back a few months.

I have been quite critical of the behavior of 8 posters on this forum.
-- Massive numbers of one-liner posts that add nothing any discussion.
-- Massive numbers of garbage posts with images and cartoons
-- A concerted effort by about 6 posters, including government employees, to blow up and shut down any negative discussion about the Baltimore Key Bridge. Note that they engineered the locking of three threads about the bridge.

How about a forum rule against those far greater problems?

Beltway, I'll say it in the nicest way possible: if eight people are calling you out, then you should reflect on yourself and think that maybe there is something you should change about yourself.

Given that the Key Bridge threads were publicly readable and tied to a major national event, the lurker‑to‑poster ratio was almost certainly enormous -- easily dozens to one, and likely into the hundreds. The visible replies in a thread don't represent the full audience, especially when hundreds of non‑members read without posting. So an "8 vs 1" headcount doesn't tell us anything about accuracy or honesty; it only reflects who chose to speak.

[freebrickproductions]

I am already having discussions with the owner of this forum, and have been going back a few months.

I have been quite critical of the behavior of 8 posters on this forum.
-- Massive numbers of one-liner posts that add nothing any discussion.
-- Massive numbers of garbage posts with images and cartoons
-- A concerted effort by about 6 posters, including government employees, to blow up and shut down any negative discussion about the Baltimore Key Bridge. Note that they engineered the locking of three threads about the bridge.

How about a forum rule against those far greater problems?

Beltway, I'll say it in the nicest way possible: if eight people are calling you out, then you should reflect on yourself and think that maybe there is something you should change about yourself.

Given that the Key Bridge threads were publicly readable and tied to a major national event, the lurker‑to‑poster ratio was almost certainly enormous -- easily dozens to one, and likely into the hundreds. The visible replies in a thread don't represent the full audience, especially when hundreds of non‑members read without posting. So an "8 vs 1" headcount doesn't tell us anything about accuracy or honesty; it only reflects who chose to speak.

If more people agreed with you, then they should say so. The people not replying to the thread could just as likely to agree with the eight or so folks callin' you out for sayin' BS.

Claimin' (or at least implyin') that a "silent majority" agrees with you is very much a fallacy, and is also almost certainly not true.