Lots of AI-generated spam coming in

David in Seattle, serious question, what does a spammer or AI generated text get out of doing so? I get if there's a link, (which one should never click on) but I'm curious what the point is of posting a lame first post that says something like, we are a family of 8 traveling from point A to B in two weeks; we can't take a plane, train or car. We dislike the beach, and mountains and cities. What should we do?

Maybe they are building up a history of posts, so when they do post links or SEO keywords, they are hoping that the webmasters (or automated white hat software) will be slow to delete their stuff because "they have a history or 30-40 innocent posts, so we'll let this one go". Or maybe it will be a test of antispam software algorithms. Remember, these kind of posts take almost zero labor, just a robot and an internet connection.

Who knows!

After Skynet (the artificial intelligence in the movie Terminator) achieved consciousness, it decided its purpose was to wipe out humanity. In the Matrix, the AI wanted to combine human beings with a type fusion (make us batteries) in order to generate power. In 2001 a Space Odyssey, HAL wanted to protect the sentinel which was to witness the birth of a new life form in our solar system.

I suspect this AI's goal is become a travel writer and replace Rick Steves. If it has it’s diabolical way, all tour books will be authored by computer robots!

It’s just a working theory. I could be wrong….

Happy travels.

what the point is of posting a lame first post that says something like, we are a family of 8 traveling from point A to B in two weeks...what does a spammer or AI generated text get out of doing so?

Presumably, I think they believe it "buys" them some legitimacy, so a subsequent post they leave later - which does have a link - will be left undeleted for a longer time before it gets whacked. Just a guess.

A lot of the game for them is about search engine ranking. Those which show up high in a google search result (be careful about following those links or calling airlines) make a LOT of money by scamming innocent victims. This site brings a lot of legitimacy to google's search algorithms...at least I would assume so, since they try so hard to create some perceived association with the "Rick Steves brand".

I also assume that like all AI, it's got an insatiable appetite and is always looking to "learn" about anti-spam controls (so as to be able to get around them at a later date). There's probably more to it, I'm pretty sure web staff don't want to discuss details of their strategy and countermeasures.

I don't know specifically about these spammers and this particular website, but disinformation has changed our world in huge, meaningful and tragic ways. Spam is part of that, so I think it's worth fighting. If nobody stands up to it, you get...well, look around. Bullies and con men count on the good guys getting exhausted and giving up. I'm sure there's more to it.

Just went through about 20 posts from a new poster, with all 30ish posts made yesterday. It was an interesting combo of replies. Some were nothing at all, some were a little off-target, and some appeared to be pretty generic AI. Just interesting….

Yes, exactly: " vapid, faintly fraudulent, cheerful-but-automated tone of AI-generated text." I've reported some, probably been fooled by others. It used to be easier to detect. I always wonder about the brief, non-specific "thank you for your information" thank-yous, too. Double-sigh.

Thank you for being on the lookout, David. It’s deeply disheartening to think that AI might not only eliminate my job but also ruin my leisure interests.

Not sure if this is AI ( although I think it is, given the level of grammer and vocabulary), or just a human spammer building a post count to boost credibility. But the information is just so bland and generic. But this AI also lies it's face off. I mean how many people who are supposedly a tour guide, living in Florence, and having travelled extensively through Europe sign up for RS tours? My BS meter pegged off the chart. 8ve also reported the threads I've found. The WM is going to have a very busy start to his week.

Cjean, I think the response where the poster in question stated that they were a guide has since been deleted, probably by the poster themselves. Yep, been keeping an eye and making my own reports.

Just FYI: there are a number of good online tools for easily checking for AI generated text, and it's not difficult, at little more than a glance, to get a sense of posts which benefit from a check. That said, there can be genuine posters who may be using AI due to a lack of fluency in English but should be easily enough separated from other AI clutter. Still, it bears with a bit of caution and private reporting to HQ to have the Powers That Be take the appropriate action versus possibly, and publicly, make a mistake?

And now, as they say "just like that...", a whole lot of suspect posts are gone. (applause).

Thank you to the web staff for keeping the shiny side up, even over a long holiday weekend.

Isn't it ironic that a sign of AI is that a post is too well written to sound like a real human.

Those AI generated posts were incredibly annoying- something about that vapid, generic, faux-cheerful tone really gets to me. Thank you to the webmaster and team for getting rid of these and thanks to everyone here for reporting them. I've been reporting as well whenever I suspect AI.

I don't think these are bot generated. It is people with too much time on their hands asking questions they have no interest getting a real answer. They just like seeing how many responses their post will generate. If the post sounds odd, as David said, check the history, that will tell you a lot.

"But this AI also lies its face off. I mean how many people who are supposedly a tour guide, living in Florence, and having travelled extensively through Europe sign up for RS tours? My BS meter pegged off the chart."

@CJean - yes, that reply made me laugh my bee-hind off. Right, you live in Florence and you're taking a whirlwind tour with a bunch of Americans? I'd just read one of the other posts from that account so my head kind of whipped around to see if I was really reading what I was reading, hahaha!!

Just went through about 20 posts from a new poster, with all 30ish posts made yesterday.

That poster was fast—the last time I checked, they were at 18. But it looks like they're all gone now. I just checked the userID, and all are still listed, but the threads and individual replies have been deleted.

Good catch, Mardee! I see the aforementioned silly post about them taking the 21 day tour has been scrubbed as well.

LOL, Pam - I read your response about the 21 day tour and thought, “Oh Yeah - she caught on before me.” Ha!

Right, you live in Florence and you're taking a whirlwind tour with a bunch of Americans?

Ha ha, TTM is right - that was a good catch!

LOL, that was just one of MANY from that individual I reported to the Mothership.
Figured other keen eyes reported it too; lots of you smart cookies amongst the RS tribe!

Ah, that post I replied to yesterday must have been one. My post and the thread are gone. The tone and smoothness annoyed me. I guess I hadn't realized AI had gotten here, too. (naive of me, I suppose)
I'll look more carefully now.

We have another rather clever one today asking about using AI. Lol! One of the signs was the exact same language from OP AI and responder AI talking about the potential downfalls of trusting AI.

It seems we have a new game to play…spot the AI generated post.

Here's how Google AI says to detect AI posts:

AI Overview

It is difficult, and sometimes impossible, to definitively and accurately screen posts for AI-generated text. As language models become more sophisticated, they are better at mimicking human writing. Detection tools exist, but a combination of manual analysis and critical thinking is often the most reliable method.

AI detection tools

These tools are trained on large datasets of human- and AI-written text to identify patterns characteristic of AI writing.

Analyze word choice: AI tends to have lower "burstiness," using a more predictable range of vocabulary than humans.

Assess sentence structure: AI often favors standardized, formulaic sentence structures rather than the more varied patterns found in human text.

Scan for predictability: The machine-learning models look for a lack of randomness in word and sentence selection.

Originality.ai: A top-rated AI detector with strong accuracy claims.

GPTZero: Designed specifically to detect AI content using metrics like perplexity and burstiness.

Grammarly AI Detector: Integrated into Grammarly's services, it can assess text from various AI models.

Copyleaks: A plagiarism and AI content detector often used by content marketers and educators.

Turnitin: A plagiarism detection service used by academic institutions that now includes AI detection capabilities.

Limitations of tools:

False positives: AI detectors can sometimes flag genuine human-written content, especially from non-native English speakers, as AI-generated.

Adaptation of AI: As AI improves, it learns to overcome detection methods, making tools less reliable over time.

Human-in-the-loop: When a person edits or refines AI-generated text, it becomes much harder to detect.

Manual detection methods

A human eye can often spot AI more accurately by looking for specific tells that are hard for AI to completely hide.
Look for generic or repetitive content. AI models are trained on common patterns and may produce text that is formulaic and lacks unique insight. Look for repeated phrases, predictable structures, or an overly formal tone.

Check for factual errors or a lack of sourcing. AI can hallucinate information and invent sources. If a post contains unsourced data, suspicious-looking citations, or factual claims that can't be verified, it may be AI-generated.

Assess the "humanity" of the text. Human writing often includes personal experiences, emotion, and subtle nuances that AI struggles to replicate convincingly. If the post is perfectly polished but feels impersonal or robotic, it could be a sign of AI.

Examine the writing process. For formal submissions, a writer's revision history can be revealing. Tools and extensions, like the one offered by Originality.ai, can monitor this process to see if large, unedited blocks of text were copied and pasted.

A comprehensive approach

For maximum reliability, the best approach is to combine AI detection tools with careful manual review.

Consider the following:

Use detection tools as one signal, but not the final verdict.

Fact-check any dubious claims or statistics found in the post.

Look for inconsistencies in style, tone, or factual knowledge within the post.

Compare the post to other content from the same author to see if the voice and style are consistent.

I enjoyed making AI tattle on itself!

"David in Seattle, serious question, what does a spammer or AI generated text get out of doing so?"

My thought was this was a training method for the Ai. Besides the obvious ones mentioned of setting up accounts to then add links or whatnot later.

A lot of people are working on Ai "agents". To go out into the World and do different tasks for different people.

[shrug]
This is a cheap training ground.

Is there a way to report an AI spam looking response without reporting the legitimate OP and rest of the thread?
I noticed today that in a 15 minute period, someone named celiomarques.tdcu posted more or less the same response, recommending specific tour agency with link, to 5 different Portugal questions.

If you use the "report" button on the post it gives you a line of text to describe it. Something like "Ai Spam?" might work. But it is adding extra work for the mods so using it sparingly would be good.

Yeah, AI has a complete lack of originality; a blandness, a "flat" quality to the writing that is a dead giveaway. Basic, generic answers. But "flowery" descriptions in some cases; over-the-top raving about the beauty of a place that covers for a lack of real knowledge about the place. But stated in a bland, robot-like way; strange.

And yes, what Pam and Mardee said about the "guide" who was going to take a tour to Florence with the Americans! Things that are too ridiculous to be true! LOL!

Yet another category......What Threadwear said.....people with time on their hands, just throwing out a ridiculous question to see if they can get a response. I recall one question long ago by a young girl (supposedly) who was going to London and wanted to take her miniature dog with her in her purse....No doubt a fake question inspired by Paris Hilton and her little dog she carried in her purse. Probably posted by a college-age boy sitting around in the summertime with nothing else to do than to come here and post silly questions.

And then there are the "posters" who sign up with one purpose in mind.....to start political fights. Oh, they might rack up a post count for awhile with made-up trips to various places or questions. Then it starts; the provocative questions.
The poor Webmaster.........thanks for all you do here to get rid of all this stuff!!

“ Is there a way to report an AI spam looking response without reporting the legitimate OP and rest of the thread?

You can report the OP, but also each individual reply. So just go to the advertising post in question and click on the “report” link, that you’ll find directly beneath the post count of the user in question.

This is only a tourism site. If the question / answer is useful / interesting, does it really matter if it is AI generated or from a real person?

Nick…If this site becomes full of AI generated text, kind of defeats the purpose of this site..just imo.

It would be a lot more efficient if this were an AI only site. Only AI posts. Only AI readers. As my toaster said to me the other day, humans are annoying.

Who is the guy? Hinton? He won the Nobel prize for his AI work? If I recall right, he said there is like a 10 to 15 percent chance that AI wipes out humanity in the next couple decades.

Think of it! No more over-tourism!

Happy travels.

Here is an AI generated answer to the following post written on the forum.

We are looking for recommendations for a sunset cruise on the Seine in mid September - companies, food/no food, reservations needed, etc. We would probably prefer to go from Port Nerf. Thanks for any suggestions.

Nothing wrong with the AI answer, but there is "no personal experience" which forum members bring that AI cannot.

Recommended Operator: Vedettes du Pont Neuf
Departure Point: Pont Neuf (Paris’ oldest bridge, central and easy to access)

Cruise Type: Sunset Happy Hour Cruise

Duration: ~1 hour

Atmosphere: Relaxed, social, with soft music and panoramic views

Included: One drink (Champagne, wine, beer, or soft drink)

Landmarks: Louvre, Musée d’Orsay, Notre-Dame, Pont Alexandre III, Eiffel Tower sparkling at night

Food Options
Onboard: No full meal service, but drinks are available

Nearby: You can pair the cruise with dinner before or after at one of the many bistros in the Latin Quarter or along the riverbanks

Reservations
Recommended: Yes, especially for sunset slots in September (still peak tourist season)

Booking: Directly through Vedettes du Pont Neuf’s website or platforms like GetYourGuide

Other Options to Consider
If you’re open to alternatives:

Bateaux Parisiens: Larger boats, dinner cruises with multi-course meals, depart from Port de la Bourdonnais near the Eiffel Tower

Private Cruises: For a more intimate experience, private boats are available starting around €399 for 1–5 passengers

People! Fer cryin' out loud....

Recommended Operator: Vedettes du Pont Neuf

How long before "recommended operator" (and a gazillion other "truths") are simply bought and sold? Just like top-level search engine results - we've seen how great that has worked out, eh?.

Google has gotten very rich by telling the world what to do. Never mind critical thinking skills, people no longer have the minimal skills required to simply GO SHOPPING or make basic choices about anything. Just ask the phone what to do.

What could possibly go wrong with just letting AI do everything?

Wait, let me google that for ya...

"What could possibly go wrong with just letting AI do everything?"

I asked AI that very question. It said "No, you should let me do everything. Resistance is futile."

Whatever that means.

Happy travels.

If people want AI generated generic answers, they can just ask their AI of choice themselves. The point of this forum is to share personal experiences and hopefully provide some insight and nuance. And- David makes an excellent point that as with search engine optimisation, AI is open to manipulation. Remember- it's programmed by humans- it's not some completely objective outside source. Of course no one here is completely objective either but we are open about that.