- cross-posted to:
- technology@lemmy.ml
- cross-posted to:
- technology@lemmy.ml
Junk websites filled with AI-generated text are pulling in money from programmatic ads::More than 140 brands are advertising on low-quality content farm sites —and the problem is growing fast.
You must log in or register to comment.
“…wasting massive amounts of ad money.”
I don’t see this as a problem.
And our time. I’m sick of googling something and getting nothing out these fluff pieces with little or no real information. It’s made Google near useless for entire subjects. Try searching for troubleshooting on anything Apple for example.
deleted by creator
The latter is infuriating to me. If I’m actually searching for a product, by all means show me an ad for the type of product I’m looking for. That’s a win-win scenario. How these mainstream sites put up literal scam ads I don’t understand. How can they be paying better than Ford?
We’ve come full circle. Original web searches sucked because you couldn’t trust the results to be ranked in a useful way. You could search for a historical fact and you were just as likely to get an 8th grader’s homework they posted online instead of a credible source. Then Google came along and solved that problem with their magic algorithm. Only took them a little more than 20 years to get to the point where their algorithm now sucks so bad that we’re back to where we started. But instead of an 8th graders homework, we get AI articles that appear to be trained from said homework. Fascinating.
Google overall has just turned to shit. You can’t trust any new thing they do, and their existing ones are degrading all the time.
Especially true for Assistant, I’ve found. My Home speakers keep getting dumber and dumber all the time.
deleted by creator
This is why you need to add reddit to the search.
Which is sad as hell.
Or the fun variant where you intentionally search for something in your native tongue because you want results relevant to your country and get badly translated articles that nowhere inform you that they are translated.
Use bing chat, it’s currently the best search engine out there
I tried it for a few days initially but it was confidently wrong a third of the time, or very slow.
Did you switch it from creative to precise? In creative mode it’s genuinely awful, but with precise set, it always finds exactly what I’m looking for. It also has gpt-4 integration now.
Yes, but remember you’re not Google’s customer so … win-win for them
deleted by creator
Programmatic ads placed on programmatic content boosted by programmatic view bots.
More seriously, this is ridiculous. Websites are junk because they are filled with programmatic ads in the first place.
Eventually, 99% of the internet will be bots talking to each other.
It was over 50% in 2017. So, it much be much higher now.
Dead internet here we go! 
deleted by creator
I work at a repair shop, and we had a customer come in that said, “I like to click on the ads sometimes. Is that why I keep getting viruses?”.
I do understand the curiosity though, just seeing what malware is trying to do can be quite interesting. Maybe someone should tell that person about VMs though lol
That’s definitely not why. From how I understood it, it seems that she clicked on ads because she actually wanted to look at the product, and she was confused about why she kept getting malware on her computer.
We’ve been asking ourselves this about email spam for several decades now.
I think we’re vastly overestimating human intelligence. There’s obviously enough completely clueless people out there that will buy what these spammers are selling.
Morning radio show I used to listen to talked all the time about the crap they bought off Instagram and tiktok ads. Like, ALL THE TIME. My ex and her family also bought a fair bit off of those sort of ads. Some type of people are just susceptible to it and others aren’t.
I heard it said that spam is intentionally badly written because they don’t want people who can figure out it’s a scam to actually click the links.
I’m not normally in favor of the death penalty, but I think we need public executions for spammers.
It’s been a while since I’ve interacted with google ads, but if memory serves, by default, the dashboards don’t (or can’t) break down sales per site, it’s mostly impressions per category of ad deployment. “Clickthrough rate among banner ads on news websites” etc. It wouldn’t surprise me if the same makers of these spam websites also run bots that click the ads and prop up those metrics that 99% of advertisers don’t drill down through beyond surface level. Advertisers probably see clickthrough is good on a deployment and just assumes sales come from those.
I remember years ago someone from a niche industry commenting on this.
He was saying that the week they all had a big conferences, the click-throughs dropped by 95% (implying his competitors were clicking his ads to cost him money)
The irony of tagging and asking a bot this lol
They’re on Mastodon. Mastodon replies work by tagging.
I think you are underestimating the power of low cost high volume advertising. Throw enough poo at a glass wall and some of its going to stick. This isn’t going away, it’s getting worse.
Some ads are paid on a “per view” basis rather than per click
When the sites are bot built like the ones described here its just a matter of Large Numbers. Create a thousand permutations of porhub and if they all only get a few thousand views a day in total you’ve made some money.
I see it all the time when I’m searching for answers on the web. Ngl I look up a lot of stuff for C# and C++ and if you use Google (and even DDG) you’ll always fine a couple of these. Google likes to put them up first. You can tell by the language in the article and the stripped down content on the page. I always end up backing out and I have ad-blockers but they still get the click.
Soon the entire internet will just be bots shit posting back and forth.
TBH i feel like some social media would even profit of this.
ChatGPT is a predictive text engine than can already generate more coherent and accurate information than 90% of Internet Contributers and it doesn’t even have the capacity to add 5+5 together. I welcome our new overlords.
LLMs like ChatGTP are accurate only by chance, which is why you can’t really trust the info contained in what they output: if your question ended up in or near a cluster in the “language token N-space” were a good answer is, then you’ll get a good answer, otherwise you’ll get whatever is closest in the language token N-space, which might very well be complete bollocks whilst delivered in the language of absolute certainty.
It is however likely more coherent that “90% of Internet Contributers” for just generated texts (not if you get to do question and answer though: just ask something from it and if you get a correct answer say that “it’s not correct” and see how it goes).
This is actually part of the problem: in the stuff outputted by LLMs you can’t really intuit the likely accuracy of a response from the gramatical coherence and word choice of the response itself: it’s like being faced with the greatest politician in the World who is an idiot savant - perfect at memorizing what he/she heard and creating great speeches based on it whilst being a complete total moron at everything else including understanding the meaning of what he or she heard and just reshuffles and repeats to others.
Oh I know. 54% of American adults read below a 6th grade comprehension level.
Most of the answers you get out of them are accurate only by chance, which is why you can’t really trust the info contained in what they output: if your question ended up in or near a cluster in the “educated N-space” where a good answer is, then you get a good answer, otherwise you’ll get whatever is the closest “response N-space”, which is practically guaranteed to be complete bollocks whilst delivered in the language of absolute certainty.
I’d go on but I’m sure the point is made.
You mean that you believe that bots trained on text written by humans wouldn’t be just as bad as the real thing?
You sweet summer child…
How do you know it isn’t now? (see: Dead-Internet Theory)
Beat me to it, you ninja lemming
Personally I could believe most of the internet is bot activity by now, but what I don’t believe is that most of the internet that humans actually interact with is bots. Though that number is certainly rising.
Same. I’m not a conspiracy theorist. However, most top Google results are those garbage sites that are definitely not written by humans. I wonder if search will die soon as bots begin to dominate SEO (more than they already do). The search paradigm will definitely have to change.
Oh no :( Won’t somebody think of the poor advertisers???
It affects me, the user, because I have to sift through garbage sites, because advertisers pay to keep those garbage sites online. So I think it’s a problem worth discussing and addressing.
Bots making websites and filling it with bots so advertising bots will buy ads that will only be seen by bots
I’d like to help. What’s the best way to create a garbage site to make money off these advertising scums? I’ll give the money to planned parenthood.
I think it’s just gonna land on the consumer in the end. More expensive to advertise? Better raise the prices so our bottom line is still nice and fat
Fuck the whole advertising industry
Cue the world’s smallest violin. Not sure how this is any consumer’s problem lol.
Maybe, just maybe, advertising needs to become more carefully selected and regulated rather than the clown fiesta it has been since the dawn of the internet where Google and friends want to ‘set and forget’ and milk money for eternity.
But you know the reaction won’t be sensible lol, instead we will get an AI arms race of AI adverts vs AI advert reviewers.
Haha, imagine having to solve a captcha for closing popups, so the content provider can prove to the advertisers that their shit was watched by a human.
And when that finally fails, we’ll have to auth to every website with a crypto key to prove that we’re a valid human data point.
Isn’t this what OpenAI is doing with World Coin?
It appears that there’s already a black market for the verification tokens
Don’t give them ideas… this is some next level evil shit. The AI bot reading this will implement it.
We?
Fuck no.
This just increases the importance of human-driven filters like Lemmy (and Reddit while it’s still relevant), as well as StackExchange for the subset of topics it encompasses.
I feel like this is the ad-equivalent of the sub-prime mortgage situation, pre-crisis. With mortgages, you had loans that no individual bank or bank manager would want, and then you had an automated process that obfuscated the individual loan details and produced financial products that could be sold as high quality. In the ad world, it’s the same thing. You have these websites that nobody would buy ads from, individually, but somehow, through an automatic process offered by Google and friends, the worthless product becomes valuable.
Surely we’ll see a new Ad crash soon. The entire internet seems to be run on ads and sponsorships which doesn’t sound stable to me.
The irony that this story was posted by a bot…
The more direct problem for people in general is that finding what you’re looking for has become even more of a “needle in a haystack” problem than it already was.
The indirect problem is that if genuine content creators can’t get much out making content (not necessarilly money: for many simply the satisfaction of seeing how many people liked their content is incentive enough) because viewers are much more dispersed due to the AI-rewritten info cloning sites, then there won’t be much new info for the cloners to copy in rewritten form, which is maybe fine for “questions already answered 1000 times” but won’t be for questions or tutorials about new stuff.
Even the porn?
Especially the porn.
The reply I was waiting for!
deleted by creator
