Author Topic: Ai moderation?  (Read 10185 times)

Offline Dadtallica

  • Gold Member
  • *****
  • Posts: 2571
Re: Ai moderation?
« Reply #60 on: October 07, 2026, 02:20:25 PM »
 

It got even better when I refined it to returned to game 2021 while the sig lines were added in 2025 as response to 2022-2024 posts (could not have it assess in-game stuff it can't access.) It kept skipping one sign line, the vox populi one so I forced it to take that into account, and it got even more betterer!!

As Gina inferred, prompts do matter and I generally keep mine (when using searches, google) as neutral as possible to avoid echo-chamber results.

Applying that same principle to the refined GPT responses/summaries and suggested prompts allow me to confidentially post this:




If anyone wants, I'll happy to share the prompts and responses. Until then, I suggest y'all review my sig lines.  :aok :noid :neener:

But then you’re tipping the scale the point of the original ask was generality and objectiveness.

I think the moral of the story is AI can absolutely moderate this place and should be probably. At the very least it could serve it up to a human with excellent context.
October Player Poll Results

Back in 2022 after a loooooong break from 2010. Old name Ratpack, same for the BBS. Old squads, Excaliburs, 172nd Rabid Dawgs.

Offline -gg-

  • Gold Member
  • *****
  • Posts: 2447
      • IceCreamOnMars
Re: Ai moderation?
« Reply #61 on: October 07, 2026, 02:27:10 PM »
But then you’re tipping the scale the point of the original ask was generality and objectiveness.

I think the moral of the story is AI can absolutely moderate this place and should be probably. At the very least it could serve it up to a human with excellent context.

What evidence do you have that Ai can "absolutely" moderate the place?  Just your feeling?

Ai has an extreme weakness for understanding certain things. It's great at analyses of data and trends and also generating content.... it's not so good at making decisions based on human interactions.

I would not want Ai moderating. I don't even like moderation to begin with, but in game - I wnt that to be a real person that does not have any bias and can be held accountable if needed.



Gina in game. Retired 11/8/26
My Band, The Jetglows
The Jetglows
Forum
Icecreamonmars.com

Offline Dadtallica

  • Gold Member
  • *****
  • Posts: 2571
Re: Ai moderation?
« Reply #62 on: October 07, 2026, 02:31:21 PM »
What evidence do you have that Ai can "absolutely" moderate the place?  Just your feeling?

Ai has an extreme weakness for understanding certain things. It's great at analyses of data and trends and also generating content.... it's not so good at making decisions based on human interactions.

I would not want Ai moderating. I don't even like moderation to begin with, but in game - I wnt that to be a real person that does not have any bias and can be held accountable if needed.

Because it’s already doing it other places and you speak to and interact with AI more than you probably realize.

Your views on AI are very anecdotal. It’s very good at making decisions. I’m not even sure what you’re referring to.

I’ll tell you what you tell me the exact question you would like me to ask about me or whatever and I’ll give you the unadulterated response?
« Last Edit: October 07, 2026, 02:36:56 PM by Dadtallica »
October Player Poll Results

Back in 2022 after a loooooong break from 2010. Old name Ratpack, same for the BBS. Old squads, Excaliburs, 172nd Rabid Dawgs.

Offline LCADolby

  • Platinum Member
  • ******
  • Posts: 7954
Re: Ai moderation?
« Reply #63 on: October 07, 2026, 02:31:49 PM »
This is joyous ChatGPT  :aok

Quote
16. The "LCADolby voice"

If I had to identify what makes a post recognizable as LCADolby, I'd say it's the combination of:

technical familiarity + compressed wording + irreverence + theatrical exaggeration + community references.

A hypothetical sentence in his style would be something like:

    "That sounds great right up until somebody actually has to fly it. Then we're all going to discover that God hates us."

That's not a quotation—just an illustration of the rhetorical pattern.

The actual posts repeatedly use the same mechanism: make a substantive observation, then wrap it in humor or exaggeration.
YouTube+Twitch+KICK - 20Dolby10

Offline Dadtallica

  • Gold Member
  • *****
  • Posts: 2571
Re: Ai moderation?
« Reply #64 on: October 07, 2026, 02:32:47 PM »
This is joyous ChatGPT  :aok

Change the personality of the model and then see what it does.

All of this is a fascinating discussion at any rate.
« Last Edit: October 07, 2026, 02:34:32 PM by Dadtallica »
October Player Poll Results

Back in 2022 after a loooooong break from 2010. Old name Ratpack, same for the BBS. Old squads, Excaliburs, 172nd Rabid Dawgs.

Offline -gg-

  • Gold Member
  • *****
  • Posts: 2447
      • IceCreamOnMars
Re: Ai moderation?
« Reply #65 on: October 07, 2026, 02:44:54 PM »
Because it’s already doing it other places and you speak to and interact with AI more than you probably realize.

Your views on AI are very anecdotal. It’s very good at making decisions. I’m not even sure what you’re referring to.

I’ll tell you what you tell me the exact question you would like me to ask about me or whatever and I’ll give you the unadulterated response?

lol. Mine are anecdotal? You just said you don't even understand how Ai prompts work.

Here's what AI has to say about an AI moderator.

_____________________________ __________________

would Ai make a good in-game moderator?



Yes—but I think AI would make a good in-game moderator only if it were used as an assistant to human moderators, not as the final authority.

For a game like Aces High, for example, AI could actually be very useful because there are a lot of repetitive things moderators have to watch for.

Where AI would be very good

1. Detecting obvious rule violations

Repeated profanity or harassment
Spam
Racist/sexually abusive language
Repeated intentional team killing
Exploiting obvious game mechanics
Botting or suspicious automated behavior
Players repeatedly evading chat restrictions

AI could look at the context, rather than simply searching for forbidden words.

For example:

"You are an idiot."

versus

"He called me an idiot because I shot him down."

A simple filter sees the same word. An AI can distinguish the situations.

2. Watching enormous amounts of activity

This is probably the biggest advantage.

A human moderator can't realistically watch 400 players simultaneously. AI can continuously monitor:

Chat
Kill/death patterns
Team switching
Collision/team-kill patterns
Suspicious movement
Repeated complaints
Player interactions
Sudden changes in behavior

It could essentially say:

Possible harassment — 87% confidence
Player X has repeatedly targeted Player Y in country and private channels for 18 minutes.
14 potentially abusive messages detected.
3 previous warnings today.

That's much more useful to a human moderator than somebody simply filing another complaint.

3. Dealing with routine stuff

AI could probably handle a lot of the mundane moderator work:

"Your message was removed because it violated the personal-attack rule."

or

"You've been muted for 5 minutes because you've continued posting after two warnings."

That would allow human moderators to concentrate on the difficult cases.

Where I would not trust AI by itself

The problem is that moderation often involves intent and context.

Imagine:

Player A: "You're cheating."

That could be an accusation.

But:

Player A: "You're cheating because you know something I don't."

Could be joking.

Or:

Player A: "You're cheating, and I'm going to report you."

Could be a legitimate complaint.

An AI can get those wrong.

And the really dangerous situation is punishment based on an AI's interpretation.

I wouldn't want:

AI confidence: 92%
Player banned permanently

without a human reviewing the evidence.

The ideal arrangement

I'd build the system more like this:

AI observes → AI analyzes → AI flags → human reviews → human decides

rather than:

AI observes → AI decides → player gets banned

And I'd give the AI access to the actual evidence that caused the flag.

For example:

MOD ALERT — HIGH PRIORITY

Player: Shane123
Possible harassment

AI assessment: 94%

Relevant chat:

Shane123: "You're a worthless..."
Shane123: "I've told you three times..."
PlayerB: "Leave me alone."
Shane123: "No, I'm going to keep..."

Previous warnings: 2
Similar incidents today: 4

Recommended action: 30-minute mute

That would be extremely powerful.

And there's another interesting possibility

AI could be better at consistency than human moderators.

Human moderators inevitably have personalities and biases. One moderator might tolerate something another moderator considers outrageous.

An AI can apply the same written rules to everybody, while still allowing the human moderator to override it.

That's particularly useful in an old-school online game where you may have a relatively small moderator staff dealing with a relatively large, sometimes argumentative community.

So my answer would be:

AI as a moderator's eyes and ears: absolutely.

AI as the final judge: probably not.

AI + human moderator: potentially much better than either one alone.

« Last Edit: October 07, 2026, 02:46:46 PM by -gg- »
Gina in game. Retired 11/8/26
My Band, The Jetglows
The Jetglows
Forum
Icecreamonmars.com

Offline LCADolby

  • Platinum Member
  • ******
  • Posts: 7954
Re: Ai moderation?
« Reply #66 on: October 07, 2026, 02:45:08 PM »
Dolby
Quote
Thank you, that was a fun read

ChatGPT
Quote
You're very welcome!  :lol😄

It was a surprisingly fun corpus to dig through. LCADolby's combination of old-school forum banter, genuine sim enthusiasm, technical tinkering, and occasional "hold my beer" combativeness makes for a much more interesting posting history than you might expect from a 7,900-post forum account.

And honestly, the continuity over 15+ years was probably the most entertaining part. The man appears to have remained remarkably committed to being LCADolby.   :lol😄
YouTube+Twitch+KICK - 20Dolby10

Offline Animl-AW

  • Platinum Member
  • ******
  • Posts: 7658
Re: Ai moderation?
« Reply #67 on: October 07, 2026, 02:52:51 PM »
That was fun, I won't post the whole LCADolby dive here, I cannot wait to read it while taking my morning dump  :banana:

Oh you read it, your ego could not resist

I predict you'll invent some way to diminish me to elevate yourself with no merit.

You're not who you think you are
AH Tech Hangar
Windows Gaming Tweaks

Offline Dadtallica

  • Gold Member
  • *****
  • Posts: 2571
Re: Ai moderation?
« Reply #68 on: October 07, 2026, 02:57:41 PM »
lol. Mine are anecdotal? You just said you don't even understand how Ai prompts work.

Here's what AI has to say about an AI moderator.

_____________________________ __________________

would Ai make a good in-game moderator?



Yes—but I think AI would make a good in-game moderator only if it were used as an assistant to human moderators, not as the final authority.

For a game like Aces High, for example, AI could actually be very useful because there are a lot of repetitive things moderators have to watch for.

Where AI would be very good

1. Detecting obvious rule violations

Repeated profanity or harassment
Spam
Racist/sexually abusive language
Repeated intentional team killing
Exploiting obvious game mechanics
Botting or suspicious automated behavior
Players repeatedly evading chat restrictions

AI could look at the context, rather than simply searching for forbidden words.

For example:

"You are an idiot."

versus

"He called me an idiot because I shot him down."

A simple filter sees the same word. An AI can distinguish the situations.

2. Watching enormous amounts of activity

This is probably the biggest advantage.

A human moderator can't realistically watch 400 players simultaneously. AI can continuously monitor:

Chat
Kill/death patterns
Team switching
Collision/team-kill patterns
Suspicious movement
Repeated complaints
Player interactions
Sudden changes in behavior

It could essentially say:

Possible harassment — 87% confidence
Player X has repeatedly targeted Player Y in country and private channels for 18 minutes.
14 potentially abusive messages detected.
3 previous warnings today.

That's much more useful to a human moderator than somebody simply filing another complaint.

3. Dealing with routine stuff

AI could probably handle a lot of the mundane moderator work:

"Your message was removed because it violated the personal-attack rule."

or

"You've been muted for 5 minutes because you've continued posting after two warnings."

That would allow human moderators to concentrate on the difficult cases.

Where I would not trust AI by itself

The problem is that moderation often involves intent and context.

Imagine:

Player A: "You're cheating."

That could be an accusation.

But:

Player A: "You're cheating because you know something I don't."

Could be joking.

Or:

Player A: "You're cheating, and I'm going to report you."

Could be a legitimate complaint.

An AI can get those wrong.

And the really dangerous situation is punishment based on an AI's interpretation.

I wouldn't want:

AI confidence: 92%
Player banned permanently

without a human reviewing the evidence.

The ideal arrangement

I'd build the system more like this:

AI observes → AI analyzes → AI flags → human reviews → human decides

rather than:

AI observes → AI decides → player gets banned

And I'd give the AI access to the actual evidence that caused the flag.

For example:

MOD ALERT — HIGH PRIORITY

Player: Shane123
Possible harassment

AI assessment: 94%

Relevant chat:

Shane123: "You're a worthless..."
Shane123: "I've told you three times..."
PlayerB: "Leave me alone."
Shane123: "No, I'm going to keep..."

Previous warnings: 2
Similar incidents today: 4

Recommended action: 30-minute mute

That would be extremely powerful.

And there's another interesting possibility

AI could be better at consistency than human moderators.

Human moderators inevitably have personalities and biases. One moderator might tolerate something another moderator considers outrageous.

An AI can apply the same written rules to everybody, while still allowing the human moderator to override it.

That's particularly useful in an old-school online game where you may have a relatively small moderator staff dealing with a relatively large, sometimes argumentative community.

So my answer would be:

AI as a moderator's eyes and ears: absolutely.

AI as the final judge: probably not.

AI + human moderator: potentially much better than either one alone.


I assure you I very much know how to prompt something I’ve made prompts longer than your DNA profile. My point was AI especially ChatGPT has evolved to the point where you don’t need to prompt it you can just talk to it.

Also, your response on moderation was exactly what I had already said lol. It absolutely can do it and serve it up to a human with great context.
October Player Poll Results

Back in 2022 after a loooooong break from 2010. Old name Ratpack, same for the BBS. Old squads, Excaliburs, 172nd Rabid Dawgs.

Offline Animl-AW

  • Platinum Member
  • ******
  • Posts: 7658
Re: Ai moderation?
« Reply #69 on: October 07, 2026, 02:58:24 PM »
This is joyous ChatGPT  :aok


edited for convenience of not owning yourself?

You word of "truce" has no value. You're the type that needs to be under contract.
« Last Edit: October 07, 2026, 03:00:38 PM by Animl-AW »
AH Tech Hangar
Windows Gaming Tweaks

Offline toasted

  • Copper Member
  • **
  • Posts: 191
Re: Ai moderation?
« Reply #70 on: October 07, 2026, 03:04:24 PM »
maybe we can start the ai moderation on the forum.

it seems like the same people always derail any serious conversation.

if your offended by this i probably mean you.
"fly as far as you can into the crash"

Offline Animl-AW

  • Platinum Member
  • ******
  • Posts: 7658
Re: Ai moderation?
« Reply #71 on: October 07, 2026, 03:09:55 PM »
maybe we can start the ai moderation on the forum.

it seems like the same people always derail any serious conversation.

if your offended by this i probably mean you.

The point of AI moderation can be helpful, but my only point is, will players trust the judgement. As I mentioned above, they already complain endlessly that text filters burned them, they blame a human mod when it was not one at all.
AH Tech Hangar
Windows Gaming Tweaks

Offline Animl-AW

  • Platinum Member
  • ******
  • Posts: 7658
Re: Ai moderation?
« Reply #72 on: October 07, 2026, 03:12:36 PM »
Dolby
ChatGPT

Post thing whole thing like I did.

Or are you just here to cherry pick self-praising?
AH Tech Hangar
Windows Gaming Tweaks

Offline -gg-

  • Gold Member
  • *****
  • Posts: 2447
      • IceCreamOnMars
Re: Ai moderation?
« Reply #73 on: October 07, 2026, 03:17:56 PM »
I assure you I very much know how to prompt something I’ve made prompts longer than your DNA profile. My point was AI especially ChatGPT has evolved to the point where you don’t need to prompt it you can just talk to it.

Also, your response on moderation was exactly what I had already said lol. It absolutely can do it and serve it up to a human with great context.

Well then you're probably just a little ignorant regarding it's capabilities.

It very much matters what you prompt it.. and to say it doesn't even need a prompt is completely ridiculous.

I'm looking at buying a 2023 Jeep Wrangler - 6 cylinder, gas, 4 door. I want to know what specific things to look for with one that has 40k miles on it.

If I don't tell AI what info I want and just type in "2023 Jeep Wrangler" it's going to describe the Jeep - not tell me anything about known issues, value or anything.

The Prompt is everything. It's not a mind reader

Gina in game. Retired 11/8/26
My Band, The Jetglows
The Jetglows
Forum
Icecreamonmars.com

Offline Animl-AW

  • Platinum Member
  • ******
  • Posts: 7658
Re: Ai moderation?
« Reply #74 on: October 07, 2026, 03:22:48 PM »
Dolby
ChatGPT

Here lil Buddy let me help you out

If you mean what LCADolby is like as a poster/personality on the HiTech Creations (Aces High) BBS, the impression from his posts is pretty distinctive:

    Very active and deeply embedded in the community. His account has thousands of posts, and he remains a regular contributor in 2026. The BBS currently lists him posting in General Discussion, the O' Club, and other forums. HHitech Creations Forums+1
    Blunt and sarcastic. He frequently uses teasing, dry humor, and deliberately provocative one-liners. For example, when someone complained about another player, Dolby replied, “Are you cereal?” HHitech Creations Forums
    Confident about his opinions. In arguments, he generally doesn't hedge much. In a 2023 discussion about alleged cheating, he compared the controversy surrounding “The Judge” to the earlier Skyyr controversy and openly discussed his own assessment of players. HHitech Creations Forums
    A serious longtime Aces High player. A 2026 stats discussion credited the Dolby identity with hundreds of kills in several tours and initially attributed 4,760 hours to him; Dolby clarified that some of those hours belonged to an earlier player who had used the name before him. HHitech Creations Forums
    Can be combative when someone rubs him the wrong way. A 2025 exchange with another poster escalated considerably, with Dolby criticizing the person's behavior and character. HHitech Creations Forums
    But he's not simply a troll. He also participates in ordinary game discussions, provides opinions on sounds/skins/game mechanics, and has been involved in the community for years. HHitech Creations Forums+1

Overall: I'd characterize LCADolby as a highly experienced, outspoken, sarcastic veteran poster who enjoys banter and isn't afraid of confrontation. He seems to occupy the role of one of those long-standing BBS personalities that people either find entertaining and knowledgeable or abrasive, depending on the interaction.

One interesting wrinkle: “LCADolby” and the historical “DoLbY” aren't necessarily the same player. Dolby himself says an earlier P-51 player used the name before him and later changed to Rogent; he says his own use of the name begins around the date his current BBS account joined. HHitech Creations Forums

If you're asking because you want to know whether LCADolby is generally respected, disliked, or considered a good Aces High pilot, I can dig specifically into that reputation.

es. Looking specifically at LCADolby's reputation as an Aces High pilot, rather than his BBS personality, I'd put it this way:
As a pilot: generally respected

There is fairly strong evidence that he's considered a legitimate, experienced, above-average pilot, rather than merely someone with a big BBS presence.

A 2026 BBS analysis of historical tour statistics put Dolby among a group of notable/high-volume pilots including Skyyr, Grizz, TonyJoey, Bruv119, 2cmex, Violator and others. Dolby's best listed tours included 503 kills in tour 168, 474 in tour 152, and 471 in tour 151. HHitech Creations Forums

That's significant because the comparison wasn't simply “people who post a lot”; it was based on actual in-game statistics.
He's not generally regarded as one of the absolute monsters

There is an important distinction, though.

If you're asking whether LCADolby is regarded as elite/top-tier in the historical Aces High pantheon, I'd say probably not at the very top.

The same statistical discussion puts pilots such as Skyyr, Grizz, TonyJoey, Shane, Kappa, Bruv119, 2cmex and Violator ahead of Dolby in various peak-tour measures. Dolby's 503-kill tour is very good, but the truly extraordinary historical numbers were substantially higher—Skyyr, for example, had 2,360 kills in one tour. HHitech Creations Forums

So I'd characterize him more as a highly competent veteran / dangerous fighter pilot, rather than “one of the handful of statistically dominant pilots the game has ever produced.”
The interesting part: his reputation is somewhat polarizing

This is where pilot reputation and BBS reputation diverge.

He's clearly been around the game and community for a very long time. His BBS account has thousands of posts, and he has been associated with LCA for many years. An old LCA player returning in 2024 specifically remembered the LCA group and Dolby was still part of that community. HHitech Creations Forums

But he's also very willing to get into arguments, and some of the criticism directed at him isn't about his flying ability at all. In a 2025 thread, for example, another long-time poster accused him of turning discussions into “drama” and called his behavior childish. Dolby, meanwhile, was openly trading insults with the other participant. HHitech Creations Forums

So there appears to be a pretty recognizable split:

    People who value his experience and combat ability: tend to see him as a knowledgeable old-school AH player.
    People who dislike his BBS style: tend to see him as argumentative, abrasive, or drama-prone.
    His actual flying reputation: considerably better than the impression you might get from some of his BBS fights.

One other thing I'd be careful about

The name Dolby has some historical-statistics baggage. In the 2026 stats thread, LCADolby specifically corrected the record and explained that some of the lifetime statistics attributed to “Dolby” actually belonged to a previous player who used the name DoLbY before him. HHitech Creations Forums

So I wouldn't take every historical “Dolby” statistic you find and automatically attribute it to today's LCADolby.

My overall read: good veteran pilot, probably respected by people who know his flying, but personally polarizing on the BBS. I wouldn't put him in the “legendary statistical killer” tier based on what I've found, but I also wouldn't characterize him as an average or mediocre pilot.

If you want, I can go one level deeper and look specifically for what other Aces High pilots have actually said about fighting against Dolby—that would give a much better picture of whether people considered him dangerous in the air.
AH Tech Hangar
Windows Gaming Tweaks