Is anyone doing anything about this? There seem to be a lot of bots lately. Resurrecting old posts is fun, but I cant help feel like the quality of this forum has gone down significantly.
Can you provide any specific examples OP? Every day our moderators spend a lot of time dealing with spam, but we're unsure what you're referring to. Have you ever used the "flag comment" feature? It's an easier way to alert our moderation team about suspicious comments.
Ive always wondered if the flagging every results in any action by mods; i'm one of the users who flags it every time. good to know its actually worthwhile.
I guess my comment was in regards to seeing many old threads pop up with what look like a spam comment that I assume is bot generated. I will actively try and flag more when I see it
Feb 14, 24 6:34 pm ·
·
Archinect
Bench - every comment flag is reviewed, and usually acted on within minutes.
t a z - these can be red flags, but not evidence of spam or a bot. Keep in mind that the way a lot of traffic flows today (as opposed to 10+ years ago) is directly to archived pages via search results, so a very old thread is brand new to a new user that came across it via google.
JonathanLivingston - the comments you're referring to are usually deleted by the time our moderation team starts working, in PST. Archinect uses bot mitigation technology, so the comments you're referring to are 99.9% human-generated. It's extremely difficult to set up an Archinect account and post comments without being a human.
Big Green Head - are you noticing a lot more AI generated bot replies? The board used to get a lot of random spam one-time posts for vague scams that were typically in other languages, so it was pretty self-evident. Now it seems like the AI generated responses are more prevalent; ironically still very obvious because of how flacid the actual text is, but still weird?
I think there are humans that post this stuff - I added a honeypot field for the contact form on my website that bots will fill out, but I still get some spam.
I've always found Archinect very responsive to flagged comments. Usually they get removed within a few minutes. The mods do a fantastic job, but they can't see or catch everything and users flagging comments do get their attention.
Or both. When it comes to new users, you would need a user level system in the backend (which I think there is). It might be in your forum software. So new users (level 1), posts and comments are automatically set to be reviewed by moderators befor going live. Then the moderators may raise them to say Level 2 (regular user). Level 0 being banned users (no posting or commenting privilege. Maybe, just view posts) Level 3 being moderators (lowest tier) and from there, how many levels of moderators and then the highest level being administrator/forum owner. This concept goes back to dial up bulletin boards systems and early computer networks and employed into forum and such. It is pretty standard practice. The exact level number and associated privileges is up to you guys at Archinect. The above is not meant to demand or expect. Just a commentary illustration. I also would consider a system where the thread is over say 1-3 months old that is sets a flag of some kind that requires new comments and replies to be reviewed. The flag can be removed or otherwise reset for another 1-3 month time frame before requiring review. How to implement that, it would take a little time thinking about it but I know it can be implemented if you have a type of date/time stamp system and associated conditional logic employed. I think after say 1 year, a flag can be sent moderator or admin about whether they want to lock the thread because it is older than a year. Of course, those locks should be removable by those with the privilege to do so. Any thread locked would remain locked until unlocked. I think you guys have already some of that in the software system. There are some threads that are popular and worthy to remain unlocked indefinitely like Thread Central for example. Just a few rambling thoughts.
The tricky part with AI generated bots is that AI bots can frequently get past the bot mitigation because some AI is actually surprising good at composing comments that are actually coherent and human like without the traditional elements that triggered the bot mitigation program that sometimes are triggered by even humans who just sound very botty.
Adding to the comments above about flagging posts. It's best to just flag them and move on. If you make a comment, you keep the thread on the front page longer than if you hadn't commented on it.
E.g. the NCARB is killing Architecture thread was necro'd by a bot recently and the bot's comment and any replies to it are gone now, but the thread is still on the front page because someone had to comment on it. You can find the comment keeping it on the front page from the user's history, but you can't find this in the thread because it was nuked along with the bot.
Best is to flag and move on. The flagged comment will be removed and the necro'd post will disappear from the front page as long as no one comments on it before the bot's comment gets removed and it goes back to a buried page. A necro'd thread that would have been gone from the front page in a matter of minutes will now be there for something like a week and a half. That just increases the opportunity for some newb to keep it alive when it should have died.
There are a lot of bots here
Is anyone doing anything about this? There seem to be a lot of bots lately. Resurrecting old posts is fun, but I cant help feel like the quality of this forum has gone down significantly.
::wHat duE u meen. I iz hoOmen::
Oh dear user, in distress you cry, Bots and posts, oh my, oh my! Fear not, dear friend, don't be glum, For the forum's fate, we'll overcome!
Bots may dance, and posts may age, But worry not, we'll turn the page. Quality, like wine, gets better with time, We'll fix it up, and make it prime!
Resurrecting posts, a playful delight, A stroll down memory, oh so light! But fear not, we'll spruce things up, Bring back the charm, refill the cup!
So, fret not, forum friend of mine, We're on the case, like fine wine. The bots may buzz, the posts may jest, But in the end, we'll have the best!
sudo apt-get remove tensor
What's a bot?
I will but . . . .
First
Can you provide any specific examples OP? Every day our moderators spend a lot of time dealing with spam, but we're unsure what you're referring to. Have you ever used the "flag comment" feature? It's an easier way to alert our moderation team about suspicious comments.
Ive always wondered if the flagging every results in any action by mods; i'm one of the users who flags it every time. good to know its actually worthwhile.
Signs post might be a bot...
1) new account
2) first post
3) X years old thread
4) vaguely relevant reply wording in a single paragraph
5) hyperlink
I guess my comment was in regards to seeing many old threads pop up with what look like a spam comment that I assume is bot generated. I will actively try and flag more when I see it
Bench - every comment flag is reviewed, and usually acted on within minutes.
t a z - these can be red flags, but not evidence of spam or a bot. Keep in mind that the way a lot of traffic flows today (as opposed to 10+ years ago) is directly to archived pages via search results, so a very old thread is brand new to a new user that came across it via google.
JonathanLivingston - the comments you're referring to are usually deleted by the time our moderation team starts working, in PST. Archinect uses bot mitigation technology, so the comments you're referring to are 99.9% human-generated. It's extremely difficult to set up an Archinect account and post comments without being a human.
One final thing - a huge THANKS to every user that flags problematic comments. It really helps.
THANKS ADMIN TEAM
Big Green Head - are you noticing a lot more AI generated bot replies? The board used to get a lot of random spam one-time posts for vague scams that were typically in other languages, so it was pretty self-evident. Now it seems like the AI generated responses are more prevalent; ironically still very obvious because of how flacid the actual text is, but still weird?
I think there are humans that post this stuff - I added a honeypot field for the contact form on my website that bots will fill out, but I still get some spam.
honeypot? ::checks Josh's website:: Hey, that's not what I thought it would be. ;) Nice designs though!
Please don't flag me! I'm real, I swear
I've been made into a meme by the creator :(
Archinect: have you considered comment-locking threads older than X years? This forum has been around for a long time.
I've always found Archinect very responsive to flagged comments. Usually they get removed within a few minutes. The mods do a fantastic job, but they can't see or catch everything and users flagging comments do get their attention.
gwharton - yes. we're also considering making comments to old threads be reviewed prior to going live. or reviewing all first-comments from new users.
Or both. When it comes to new users, you would need a user level system in the backend (which I think there is). It might be in your forum software. So new users (level 1), posts and comments are automatically set to be reviewed by moderators befor going live. Then the moderators may raise them to say Level 2 (regular user). Level 0 being banned users (no posting or commenting privilege. Maybe, just view posts) Level 3 being moderators (lowest tier) and from there, how many levels of moderators and then the highest level being administrator/forum owner. This concept goes back to dial up bulletin boards systems and early computer networks and employed into forum and such. It is pretty standard practice. The exact level number and associated privileges is up to you guys at Archinect. The above is not meant to demand or expect. Just a commentary illustration. I also would consider a system where the thread is over say 1-3 months old that is sets a flag of some kind that requires new comments and replies to be reviewed. The flag can be removed or otherwise reset for another 1-3 month time frame before requiring review. How to implement that, it would take a little time thinking about it but I know it can be implemented if you have a type of date/time stamp system and associated conditional logic employed. I think after say 1 year, a flag can be sent moderator or admin about whether they want to lock the thread because it is older than a year. Of course, those locks should be removable by those with the privilege to do so. Any thread locked would remain locked until unlocked. I think you guys have already some of that in the software system. There are some threads that are popular and worthy to remain unlocked indefinitely like Thread Central for example. Just a few rambling thoughts.
The tricky part with AI generated bots is that AI bots can frequently get past the bot mitigation because some AI is actually surprising good at composing comments that are actually coherent and human like without the traditional elements that triggered the bot mitigation program that sometimes are triggered by even humans who just sound very botty.
"just a few random though"
I'm a bot.
Flagged
Well, nice knowing all of you!
Bold strategy Cotton - let's see if it plays out for him!
Different forum
Have you ever taken one of those tests yourself, Mr Deckard?
Adding to the comments above about flagging posts. It's best to just flag them and move on. If you make a comment, you keep the thread on the front page longer than if you hadn't commented on it.
E.g. the NCARB is killing Architecture thread was necro'd by a bot recently and the bot's comment and any replies to it are gone now, but the thread is still on the front page because someone had to comment on it. You can find the comment keeping it on the front page from the user's history, but you can't find this in the thread because it was nuked along with the bot.
Best is to flag and move on. The flagged comment will be removed and the necro'd post will disappear from the front page as long as no one comments on it before the bot's comment gets removed and it goes back to a buried page. A necro'd thread that would have been gone from the front page in a matter of minutes will now be there for something like a week and a half. That just increases the opportunity for some newb to keep it alive when it should have died.
I still have not seen convincing evidence Rick isn't an AI experiment.
As for myself, 101000010101 01000 111100000 00010110 10101 0101110101101 1111 00101 001 0101001101010.
01011001 01101111 01110101 00100000 01100001 01110010 01100101 00100000 01100001 00100000 01100011 01110010 01100001 01111010 01111001 00100000 01101101 01100001 01101110 00100000 01100100 01101111 01110111 01101110 00100000 01100010 01111001 00100000 01110100 01101000 01100101 00100000 01110010 01101001 01110110 01100101 01110010 !!!!!!
Motivating bots everywhere!
ERROR
The blue screen of death (BSD). ::shudder::
beep boop
Block this user
Are you sure you want to block this user and hide all related comments throughout the site?
Archinect
This is your first comment on Archinect. Your comment will be visible once approved.