TVB - I noticed that too. Also responding to each others posts. Less work for the rest of us.

I’m a bit torn about how to deal with the issue. As I see it, the bots are there to do low-stakes testing for language models. The goal is to fool people into thinking they’re real to prove that the algorithms are advanced enough to target bigger platforms. 

If you identify them immediately, it helps other people avoid being tricked, although it also alerts the bots that they’ve been detected and gives them feedback that the algorithms need adjustment. Theoretically, if you identify them, you’re actually helping them by addressing their weaknesses.

If you do nothing, it opens the door for more to enter and it increases the likelihood that people who don’t catch onto the bots will be fooled. The first bot in the forum recently got several genuine replies and likes before it was exposed, and in a short time it introduced several other bots.

Maybe immediately flagging and removal is a way to keep the site’s integrity, although it does mean real people might accidentally get flagged. It would be unfortunate to be put in the position of having to prove you’re real to stay in a new forum after only a post or two. Maybe that’s unavoidable.

The site is only as capable of detection as its programming allows. Given that it follows a format that long predates bots and with moderators that aren’t involved in the forum, it mostly comes down to the forum having to police itself to fend off fakery.

tonyelder - Back to your original question. I think "friend" requests are handled completely within the forum. If you get a request it shows up if you're logged in, and whether you accept or reject it, there is no more information revealed about either party than what's already on the forum.
Personal messages are different. Its been a long time, but I think you receive a notification through email. The message (or the notification) will include the sender's actual email, which is how you respond, so the sender will receive your actual email too. This can be good if you want to have further communication with the member, but not so good if you want to keep your identity private.

Hmm MorganBright and EvanTodd just both said the same exact thing, replying to eachother…
OK I guess we can all just leave now and let the bots have a nice chat about fiddling, with eachother!

Log into either Mandolin Cafe or TalkBass and you'll get tested by a "bot check" entry wall. It takes a few second extra. Probably worth it. TB actually has a list page, exposing the bots by name. How they get thru the wall, a good question?

Y'know, bot or not, I don't feel comfortable exchanging with someone with nothing entered on the About page. WTF?  I've never bought from the classifieds on any site, but if you're selling and you haven't filled anything on the About page, have a re-think.

They're getting cleverer by the minute. Both "Evan Bright" and "Morgan Todd" have updated their bios with generic information, age, etc., and even listed websites that turn out to belong to random standup comedians. (They both conveniently left out the ":" in the hyperlink so add it back in to view the misappropriated website).


I don't see why we should put up with this.

IMO, probably not bots.

But obvious fake. Most likely just low life con man... setting up some for deception, scam they plan to run. They seem be using some AI to help, but mostly using just old fashioned unsophisticated con man skills, relying social engineering aspects.

These have the similar tell-tale signs, and basic foundation used in a handful of various schemes... maybe with new twist. I would be careful with divulging too much about how can figure out they are con, as some of that can help them to hone being better at fake. It's easier to spot them if they keep on using the same poor tools methods.

Previous similar attempts seemed to just eventually went away by ignoring them. Might be a new twist though.

quote:

---

Originally posted by DougD

tonyelder - Back to your original question. I think "friend" requests are handled completely within the forum. If you get a request it shows up if you're logged in, and whether you accept or reject it, there is no more information revealed about either party than what's already on the forum.
Personal messages are different. Its been a long time, but I think you receive a notification through email. The message (or the notification) will include the sender's actual email, which is how you respond, so the sender will receive your actual email too. This can be good if you want to have further communication with the member, but not so good if you want to keep your identity private.

---

Just like messages, friend requests prompt notifications through email and you can see the address of the sender as well as the avatar.

Yup, my friend request from Grady Travis came from an email associated with "nimeshpatel.org" ...that's a bit Sus (as my kids would say).

quote:

---

Originally posted by EvanBright

Yep, definitely seems like a new wave of bots. Getting a friend request from one of them would make me suspicious too. ??

---

The irony is funny.

Karma comes around.

AI taking over fast. I guess our collective goose is finally cooked, and all that time we spent hiding under our desks from nuclear bombs we missed the oncoming, welcomed robots, descending upon us in such a friendly and fascinating-yet-blatant skulduggery.

TVB and NCnotes - Maybe things have changed - as I said its been several years. I never could see the point of "friends" and I have a bunch of unaswered requests (apologies to those few of you who're still around) and I don't think there were email notifications for any of them. Certainly there is no email address included in any of the requests, just the screen name. Maybe its different now, and that's good to know.
NCnotes - I just noticed "Grady Travis" this morning. Good choice of name, since it claims to be a guitar player!

GRady, Evan, Alex, you are all invited to my in-person jam! WHat tunes do ya play?

Also bot maker, you need to invent some women…
some of us women do fiddle, y’know….

Oh no...bots are sexist?

The latest wave is exposing the weakness in the approval process. Without detection software or active investigation by moderators before approval, it’s easy for bots and agents to set up accounts that are legitimate enough to get in the door.

The same problems have arisen with Maestronet and Violinist.com, but in both of those cases the attempts have been thwarted so far. In the former it’s handled by an active group of moderators who read posts, make edits, and remove posts when there are issues. The latter site does not have active moderation in threads, but the moderator follows everything closely and has made it clear she is committed to keeping the site as human as she can. On the Fiddle Hangout there’s more of a problem because the moderators are either vacant or uninvolved. The only way to combat infiltration as things stand is to report the posts and hope that someone in admin is going to delete them or block the fake posters.

quote:

---

Originally posted by NCnotes

GRady, Evan, Alex, you are all invited to my in-person jam! WHat tunes do ya play?

Also bot maker, you need to invent some women…
some of us women do fiddle, y’know….

---

They might be into tunes something like this.

I like how it says "LIVE" Bluegrass Music.
 

That band has been prolific in the studio in last few months, cranking out lot of recordings. Others are creating videos as well. Possibly FHO account goal is create profiles for the band bio, or followers?

There's also lots of things like this that includes that depicts female like. (would hesitate to call them women, which is associated human)

There's definitely an escalation of misinformation on YT. Not overnight, but like a matter of a couple weeks. It's just crazy. If you're looking for actual instruction on something, think twice, and use your common sense.

quote:

---

Originally posted by farmerjones

There's definitely an escalation of misinformation on YT. Not overnight, but like a matter of a couple weeks. It's just crazy. If you're looking for actual instruction on something, think twice, and use your common sense.

---

It's gotten so bad that one can't believe anything they see on Youtube.. Even fake Bigfoot videos look real... I'm a believer and even I don't believe them...Almost every Youtube topic shows a phoney AI photo..!!

I've clicked onto some YouTube videos...not music but just informative stuff or something, and right away, seems like at least half of them appear to be AI...it's not something I can say with accuracy, but something about the eyes looks eerie and unreal to me. I close out that YouTube immediately. I mean, maybe the information would be correct or helpful, but I'm just lookin' for more human-to-human info, whenever I can find it anymore.  The human-to-human content may or may not be correct or helpful, but I'm still feeling like I'd much rather listen to humans than robots.

Wow, I did not know how advanced AI music is getting. It still feels like a cheesy copy to me, though. I can’t believe they have the gall to put LIVE in the bluegrass band title?

This is a big topic, but I read that there are lawsuits by musicians who feel like they got ‘sampled’ into the algorithm. LIke if you say “Generate a song in xxxxx’s style” the AI can/will do it. Through ripping off the creations of artists (whether musicians or visual artists) the AI is enriching its developers greatly…

quote:

---

Originally posted by NCnotes

Wow, I did not know how advanced AI music is getting. It still feels like a cheesy copy to me, though. I can’t believe they have the gall to put LIVE in the bluegrass band title?

This is a big topic, but I read that there are lawsuits by musicians who feel like they got ‘sampled’ into the algorithm. LIke if you say “Generate a song in xxxxx’s style” the AI can/will do it. Through ripping off the creations of artists (whether musicians or visual artists) the AI is enriching its developers greatly…

---

You understand even clicking on some misinformation to point out the error, that's a CLICK, also. So they "win" either way. I know, it's hard to ignore. I have to stick with my "known good" subscriptions, only.

AI gonna take bigfoot's job, too !

Reminds me of the 80s...when electronic keyboards became affordable and accessible to the general public...all those synthetic sounds...fun to play around with but...

I don't see much AI slop on youtube. I think my browser is set to block slop.

The only music related youtube videos that kind of annoy me are the "spontaneous" street performer videos with multiple camera angles, near studio sound quality, with little background noise. I have had regular folks who are wowed by them send them to me...

and its awkward because I don't want to be rude and kill their joy, haha...so i say something like "thanks! quite talented!" and move on.

Like the canned talent shows like America's got talent... I can't watch that stuff. so i don't

A few months ago YouTube was flooding my algorithm with AI-generated videos of “zombie heavy metal bluegrass” and I remember reading elsewhere that AI bluegrass was suddenly becoming popular among listeners. That seems to have died down, but I’m not sure if that’s because I didn’t listen to enough to make it worth promoting to me or because the trend was short-lived.

With forums, it seems that there are waves of fake accounts that appear. Their purpose seems to be to gain entrance to forums and to see if it’s possible to go undetected. The longer one remains and the more people interact with it, the more data it can collect for training subsequent attempts.

Does anyone remember the story about the early AI program that composed daily poems? Its posts were getting lots of attention and everyone was amazed at the “cleverness” of the algorithm. It had become a viral sensation and was used as a predictor of what could eventually be accomplished by AI. But then a few posts raised eyes and an investigation was done. As it turned out, the AI had been unable to compose anything of any value, so the human running the account had been composing the poems himself daily. All the excitement was about a machine making new compositions, so once the truth was revealed, the account lost its followers for two reasons:
1) the poetry wasn’t written by an algorithm and therefore had no value as an AI predictor
2) the poetry, while better than a machine could compose, was not very good by human standards. People realized that they’d actually lowered the bar to accept machine made poetry because they were so eager to see something impressive.

I agree that it’s possible for a scammer to accomplish what’s being done here without the use of AI. It just seems that the adoption of the technology makes it more convenient to quickly manufacture multiple accounts and bake in slight variations.

Would Morgan Todd translate roughly into "Morning Death?" Somebody with that name should correct me and forgive me.

Not my name, but I think the literal German spelling would be "Morgentod," with "Tod" pronounced like "tote." I don't think its a word though.