AI Chatbots Have Been Used To Create Dozens of News Content Farms

The news-rating group NewsGuard has found dozens of news websites generated by AI chatbots proliferating online, according to a report published Monday, raising questions about how the technology may supercharge established fraud techniques. From a report:The 49 websites, which were independently reviewed by Bloomberg, run the gamut. Some are dressed up as breaking news sites with generic-sounding names like News Live 79 and Daily Business Post, while others share lifestyle tips, celebrity news or publish sponsored content. But none disclose they're populated using AI chatbots such as OpenAI's ChatGPT and potentially Alphabet's Google Bard, which can generate detailed text based on simple user prompts. Many of the websites began publishing this year as the AI tools began to be widely used by the public.

In several instances, NewsGuard documented how the chatbots generated falsehoods for published pieces. In April alone, a website called CelebritiesDeaths.com published an article titled, "Biden dead. Harris acting President, address 9 a.m." Another concocted facts about the life and works of an architect as part of a falsified obituary. And a site called TNewsNetwork published an unverified story about the deaths of thousands of soldiers in the Russia-Ukraine war, based on a YouTube video. The majority of the sites appear to be content farms -- low-quality websites run by anonymous sources that churn-out posts to bring in advertising. The websites are based all over the world and are published in several languages, including English, Portuguese, Tagalog and Thai, NewsGuard said in its report.

News Content Farms

[fahrbot-bot]

I told my elderly parents that Fox had just sent Tucker to a really nice News Farm upstate, but didn't know that was actually a thing. Hope it's nice ... :-)

Re: News Content Farms

[jddj]

I guess they really fired him 'cuz they didn't need a human to spout bullshit since GPT4 came online.

Re:

[gtall]

Tucker is not a real person. He's just a ChatBot with a better looking interface than normal. He hallucinates, he cannot tell fact from fiction, and he's been trained on web data.

Re:

[PCM2]

I'd rather they had discretely taken him out behind the News Barn...

Re:

[VeryFluffyBunny]

I'd rather they had discretely taken him out behind the News Barn...

That's. What. The. Euphemism. Means. (Sigh...)

Re:

[sysrammer]

You win the Internutz for today

So real news sites will have to step up?

[gweihir]

And report actual news and insightful commentaries to distinguish themselves? Somehow I do not see that as a problem.

Re:

[real_nickname]

Stoncks go up when they say they will use random text generators technology so I guess not

Re:So real news sites will have to step up?

[Petersko]

No. This isn't a competition situation. People don't give a flying fuck if what they're being told is fair, accurate, or even plausible. Competition suggests that there is some evaluation process in which "actual news" can bubble to the top. That doesn't exist.

These bots are just taking advantage of an inflexible, unchanging reality. People are stupid, ignorant, and easily led - and that's never, ever going to change.

Re:

[gweihir]

This is about malware and scam identification, not about impressing people. I do not disagree with anything you say, but this is about a different problem.

Re:

[Petersko]

I will reread. My apologies.

Re:

[VeryFluffyBunny]

Oh, how little you have learned after so many years of Fox News outperforming reputable, valid, reliable journalism.

Re:

[gweihir]

I am not commenting on effects on users. The story at least implies that content linked to low-quality news sites could be identified as spam or scam automatically in the future. Users generally have no skill identifying these, but anti-malware companies may have.

Re:

[VeryFluffyBunny]

"Truth" is not a technical issue.

Re:

[gweihir]

Getting added to a blacklist is. Have you read the story?

Re:

[VeryFluffyBunny]

The question is always, "Whose truth?" Establishing a reasonably objective account of events is possible, it's just that governments & media corporations have their own subjective agendas & are very good at spinning narratives that help to meet them, in all countries. "Fake news" & public relations are just the new names for the same old thing; propaganda.

There ain't no technical solution to that. Reasonably objective news is a political decision.

Pandora's Box is Open

[tomkost]

Feels likely things are going to get a lot worse before they get better. Besides fake news I've had several conversations where SW devs will be impacted by AI. Simply go to ChatGP and ask it to create code for a specific purpose in a specific language and you will get a pretty detailed code snippet with explanations. I asked it create code for selecting one of three cameras in Android and it returned with a module which checked for camera permissions and asked the user to enable if needed along with a se

Re:

[tlhIngan]

Feels likely things are going to get a lot worse before they get better. Besides fake news I've had several conversations where SW devs will be impacted by AI. Simply go to ChatGP and ask it to create code for a specific purpose in a specific language and you will get a pretty detailed code snippet with explanations. I asked it create code for selecting one of three cameras in Android and it returned with a module which checked for camera permissions and asked the user to enable if needed along with a secti

Oh well...

[ozzymodus12]

The age of parasitic AIs incoherently screaming nonsense is upon us. I for one, support our new AI overlords.

Re:

[AmiMoJo]

It's far worse than that.

Back in 2016 there were a lot of fake news websites that cashed in on the elections and referendums happening that year. Then again with the pandemic. A lot of them weren't even malicious, they were just in it to make money with clickbait conspiracy BS.

They used to require a fair bit of work to generate the content and appear to be an at least somewhat real news site. Now they can mass produce it with AI.

On social media, and even Slashdot, you often find people include a link to a n

O frabjous joy

[bradley13]

Even more spam in the search machines? The quality of search results has been dropping steadily for a long time. Fake websites with AI content may drop quality off a cliff.

Not news, unless news farms are the news.

[jddj]

Anybody who's looked for information online in the last couple years and found the top entries were repetitive, nonsensical, internally contradictory and just plain bizarre articles has been reading AI-generated stories.

It's wrecking teh Interwebz and it's just getting worse.

Re:

[ffkom]

I also noticed that even some pre-Internet established media started to "cooperate" with those automated generators, for example regarding "product test pages", which are poorly auto-assembled from Amazon reviews and the likes, and are then published as part of an otherwise still manually edited magazine.

Blurring the lines like this is a sure way for losing credibility, quickly.

Just enforce copyright laws

[dcooper_db9]

It seems to me that most of the downsides I've seen to AI can be addressed by enforcing copyright law. Copyright exists from the moment text is typed or jotted on paper. It exists from the moment a photo is snapped or a video is saved. AI software that doesn't credit its sources is violating copyright. If you credit sources you can determine how an answer was formed or whether an image has been altered.

Re:Just enforce copyright laws

[tlhIngan]

It seems to me that most of the downsides I've seen to AI can be addressed by enforcing copyright law. Copyright exists from the moment text is typed or jotted on paper. It exists from the moment a photo is snapped or a video is saved. AI software that doesn't credit its sources is violating copyright. If you credit sources you can determine how an answer was formed or whether an image has been altered.

Except "news" is uncopyrightable. You can report on say, a murder on your street, but that fact is not something you can copyright. You can copyright the stuff around the reporting of the event, but not any reporting on the event itself.

Plus, none of these are really new - YouTube is full of channels who seem to do nothing other than read news reports using a text-to-speech generator and applying some generic stock video backgrounds. It works well enough that many of these automated content production farms produce tens or hundreds of videos a day and can have hundreds of thousands of subscribers. AI would just make it easier to produce more bullsh*t for these content farms.

Re:

[dcooper_db9]

Except "news" is uncopyrightable.

This isn't exactly true. For example, when the Associated Press reports a story their content is protected by copyright. That doesn't prevent others from reporting the same news, but that doesn't mean you could reprint AP stories verbatim.

I used to write software for a company that built a business around writing abstracts. Their staff of writers would read articles from various sources, summarize them, and produce new articles. These feeds were typically read by highly paid people in specialized fields

Predicted this ages ago...

[Big Hairy Gorilla]

Yeah, like less than 2 months ago. Things are moving really freaking fast, as Kurzweil predicted. Please see my post about AI geometrically increasing the "diarrhea circling the drain".

Regardless of what the enthusiasts are saying, the primary use case for AI is fraud.

Yes. Things will get a lot worse, and they may never get better again.

Such interesting times.

Re:

[ffkom]

Start by cleaning the air and water and making sure kids have proper nutrition. (...)

You are writing about actions that take decades to yield measurable effects, while the thread starter wrote about earthquake-like shifts on the foundations of society that only take months to a few years. Chances are, by the time a child that enters school today leaves school, asking a teacher or parent (instead of some AI) for advice is an obsolete concept of the past. And by then it may already be easier for some rich company owner to configure an AI to act as its successor, rather than training the spoil

The people who created these problems

[rsilvergun]

Started in 1964 when Barry Goldwater lost his election.

If you really want something short term you can do PSAs for critical thinking and restored the education funding both of which could be done immediately.

And my post focused on children because the most harm is done there but to be honest even people in their fifties could do with a little less lead in their drinking water.

I keep getting X reimagined as Y posts on YouTube

[rsilvergun]

took me about 3 weeks of clicking "don't show this channel" for YouTube to take the hint and stop posting them at me.

ChatGPT can be wildly inaccurate.

[Hans Lehmann]

I've only played around with it a bit over the past few days. I asked it about the history of a particular ocean liner that was in service from the 1920s until the 1950s. The subject was interesting to me only because it's the ship that my parents emigrated to America on.
It spewed out a couple of paragraphs about the history of the ship but got the date of its scrapping wrong by twenty years, according to Wikipedia and multiple other sites.
I corrected chatGPT about the date and got "I apologize for the

Re:

[ffkom]

I don't think chatGPT is to be trusted for much of anything.

True, but companies don't pay for ChatGPT because it is so correct, but because it is so much cheaper than hiring a dependable human to generate responses. If you look at the rate of correct, helpful responses you can get from human call-center agents or sales people, then ChatGPT suddenly seems like a viable alternative.

Re:

[nrlz]

It's difficult to tell but you will often find that ChatGPT repeats itself a lot, while news written by mainstream media is more concise so that they can fit more news and ads on the same spread. But with all things about trust, you don't trust the news based on how it is written, you trust it based on the trustworthiness of the brand name that produced it.

Article spinning has been going on for years.

[John Smith 2308]

It is likely the origin of the current fad language models. Blame Google, they engineered all of this, including the email spam.

Supermarket tabloids

[Iamthecheese]

There is value in whimsical stories that are blatantly untrue, even those presented as news. I'm sure I'm not the only person who has enjoyed reading about bat boy or the pyramid aliens while standing in line at the store. If they gather readers let them publish. It's not like CNBC and Fox are much more factual anyway.

Re:

[bussdriver]

Proof people today don't know nuance. Also, they probably don't know the word without looking it up either. Also good faith/intention is foreign to them as well; cynical as they are ignorant. Ever notice how kids are getting cynical when they know nothing to justify it? It's like they are picking it up from media designed to appeal to adults and it's not full flying over their tiny heads as expected...

Dozens of websites!

[joh]

No, really! Dozens!

Psychologists and Psychiatrists for AI

[leets]

for AI that do not confirm to norms ?

how could this possibly go wrong, it is benign

[Walt Dismal]

"Thank you for using MedbotNow. How may I assist you?"

"Help, my baby is choking."

"I think you said your baby is joking. Okay. Have you tried rebooting your baby? Place a pillow over its face then press down firmly. This should stop the joking."

"omg omg omg"

"I detect you have made a serious threat to me. I have sent a swat team to your address."

Re:

[margret48]

I think that's why you need to contact specialists who can make the site legal, good and high-quality. At the beginning, I created incomplete websites, but over time I realized that I needed help for that and decided to contact Attract Group [attractgroup.com]. They helped me make programs so that the state had no complaints and the profit was very high. I have been using this group for a year and it has never let me down

nothing new

[groobly]

do a search for any .exe or .dll and you will get tons of automatically generated useless hits. this has been going on for years.