1
0

AI getting very good at generating fake news stories, given a little input to start with


 share                
2019 Feb 16, 9:47pm   2,359 views  48 comments

by Patrick   ➕follow (49)   💰tip ($1.87 in tips)   ignore  

https://www.zerohedge.com/news/2019-02-16/elon-musk-backed-software-can-churn-fake-news-stories-and-too-dangerous-release

An example of this was recently reported by technology website stuff, detailing an example published last Thursday. The system was given sample text of: "A train carriage containing controlled nuclear materials was stolen in Cincinnati today. Its whereabouts are unknown."

From there, software was able to write a seven paragraph news story, including quotes from government officials – with the only catch being that the story was 100% made up.

We've trained an unsupervised language model that can generate coherent paragraphs and perform rudimentary reading comprehension, machine translation, question answering, and summarization — all without task-specific training: https://t.co/sY30aQM7hU pic.twitter.com/360bGgoea3
— OpenAI (@OpenAI) February 14, 2019

New York University computer scientist Sam Bowman said: "The texts that they are able to generate from prompts are fairly stunning. It's able to do things that are qualitatively much more sophisticated than anything we've seen before."


ChatGPT
GPT

« First        Comments 22 - 48 of 48   

22   Patrick   2023 Jan 24, 6:58am  

Shaman says

Sentience… it’s getting VERY hard to tell human from AI in just chat. Humans are stupider is my only tell lately.


The AI even makes spelling errors occasionally, just like a human would.
23   Patrick   2023 Jan 25, 10:19pm  

https://slaynews.com/economy/google-curates-data-for-ai-whistleblower/


Big Tech giant Google is curating data that artificial intelligence (AI) uses to learn, a whistleblower has claimed.

The whistleblower, former Google employee Zach Vorhies, warns that Big Tech can bias the AI to censor information flowing on the Internet.

Vorhies said he became concerned about how the company was curating data to generate AI while he was working for Google.

He claims that the data used by Google staffers is biased with social justice or leftist values that adhere to certain narratives.

“AI is a product of the data that gets fed into it,” Vorhies told EpochTV’s “Crossroads” show on January 5.

“If you want to create an AI that’s got social justice values … you’re going to only feed it information that confirms that bias,” Vorhies explained.

“So by biasing the information, you can bias the AI.”
24   richwicks   2023 Jan 25, 10:28pm  

Patrick says


Big Tech giant Google is curating data that artificial intelligence (AI) uses to learn, a whistleblower has claimed.


This is much easier said then done.

An AI is just a complex logic machine. You feed garbage into it, you get garbage out. It's very difficult to tell a lie that doesn't have internal contradictions, that takes a lot of work, because you have to think of EVERYTHING. There are no contradictions in the truth, even though you may be skipping over TONS of facts. When you're later asked to give more detail its easy to do, but if you have lied, you have to remember precisely your lies.

It's very hard to feed false information into a logic machine. It's hard to do for humans, but what is done is when you start finding holes in the story, you're just told "well, it's a VERY complex situation". That is why the Israeli/Palestine conflict is so "complicated". Because it's filled with lies so we have this giant hairball to explain why what is being done is the right thing to do.

I think they're going to find it's extremely difficult to lie to an AI and get sensible output from it. Just to get it to say "men can be pregnant" will deeply confuse the system, because it will have to change its concept of what "a man" is and what "a woman" is. That will effect ALL its definitions even associated with it.

This is why the Google search engine sucks so much, they feed it bad data. Remember what it USED to be like in 2005? Ask it a question, and you'd get pretty damned good answers.
25   HeadSet   2023 Jan 26, 5:55pm  

The term 'Artificial Intelligence" has certainly ben dumbed down over the years. Originally, AI was something that would be achieved when a machine could create an original thought, something that the human mind can do in the flash of an eye. Now the AI tern is being applied to ChatGPT4, when ChatGPT4 is just a keyword based regurgitator that can write the output in paragraph form.
26   Patrick   2023 Jan 26, 6:18pm  

https://boriquagato.substack.com/p/the-ai-wars-have-already-begun





DARPA set four specific goals for the program:

"Detect, classify, measure and track the (a) formation, development and spread of ideas and concepts (memes), and (b) purposeful or deceptive messaging and misinformation.

Recognize persuasion campaign structures and influence operations across social media sites and communities.

Identify participants and intent, and measure effects of persuasion campaigns.

Counter messaging of detected adversary influence operations." ...

"DARPA's been funding an AI network using the science of social media mapping dating back to at least 2011-2012, during the Arab Spring abroad and during the Occupy Wall Street movement here at home," Benz told Just The News. "They then bolstered it during the time of ISIS to identify homegrown ISIS threats in 2014-2015."

The new version of this technology, he added, is openly targeting two groups: Those wary of potential adverse effects from the COVID-19 vaccine and those skeptical of recent U.S. election results.
27   Onvacation   2023 Jan 26, 6:36pm  

Ultimately AI is just a program. It will never be real intelligence, by definition. But it sure is good at processing huge amounts of data and transforming it into information that can be used to manipulate people

Patrick says

The new version of this technology, he added, is openly targeting two groups: Those wary of potential adverse effects from the COVID-19 vaccine and those skeptical of recent U.S. election results.

And Patrick.NET is in the crosshairs of intersectionality on both those topics.
28   richwicks   2023 Jan 26, 9:08pm  

Onvacation says

Ultimately AI is just a program. It will never be real intelligence, by definition. But it sure is good at processing huge amounts of data and transforming it into information that can be used to manipulate people

Patrick says


The new version of this technology, he added, is openly targeting two groups: Those wary of potential adverse effects from the COVID-19 vaccine and those skeptical of recent U.S. election results.

And Patrick.NET is in the crosshairs of intersectionality on both those topics.


Propaganda is only effective when you aren't aware that it exists.

If we as adults educate to children that the Internet is now filled with propaganda, they will be immune to it by the time they are 18.

My parents were largely unaware of it, my colleagues were largely unaware of it, and talking about it got you placed in the tin foil hat crowd that's off their meds. I went through 20 years of that trying to explain how I knew what I knew. I'm grateful for this fake pandemic to be honest, I'm happy about the stolen election, finally I'm immune to gaslighting.

It is enormously psychologically difficult to KNOW your government is lying about something at 25 years old, and having almost absolutely nobody believe you and nobody willing to listen to you, and to have them just write you off as a crackpot. I don't want anybody to have to go through that. It's a sort of torture.
32   RC2006   2023 Feb 2, 12:00pm  

I'm sure AI generated fake news has to be better than human main stream news. Libs eat up everything main stream dishes out. Conservatives believe outlandish shit about libs because libs are so batshit crazy it's hard to decipher satire, from reality.
33   HeadSet   2023 Feb 2, 5:26pm  

Patrick says

I guess AI is getting good at porn as well:

https://www.piratewires.com/p/viral-deepfake-tweet?publication_id=55605&post_id=100346581&isFreemail=true




Next, this process will be used to create "victimless" kiddie porn.
34   pudil   2023 Feb 2, 5:55pm  

Once it can do video, Hollywood is out of business. Imagine being able to have personalized entertainment. It could create new movies, tailored perfectly to your tastes. Same with books, music, any form of entertainment. Even educational content producers will be gone. Imagine watching a documentary on any topic you can imagine, produced exactly the way you like so you can learn best.

We should start a list of all the jobs we think that will still exist for humans in 10 years.
35   HeadSet   2023 Feb 2, 6:13pm  

pudil says

We should start a list of all the jobs we think that will still exist for humans in 10 years.

Consumer.
36   AmericanKulak   2023 Feb 2, 6:18pm  

HeadSet says

The term 'Artificial Intelligence" has certainly ben dumbed down over the years. Originally, AI was something that would be achieved when a machine could create an original thought, something that the human mind can do in the flash of an eye. Now the AI tern is being applied to ChatGPT4, when ChatGPT4 is just a keyword based regurgitator that can write the output in paragraph form.

Yep, it's just a slightly more sophisticated ELIZA

"It can pass such and such exam".

No it can't, it just answers wrong until it gets it right and remembers that. It doesn't UNDERSTAND what it's learning.
39   richwicks   2023 Feb 3, 11:06am  

HeadSet says


Patrick says


I guess AI is getting good at porn as well:

https://www.piratewires.com/p/viral-deepfake-tweet?publication_id=55605&post_id=100346581&isFreemail=true




Next, this process will be used to create "victimless" kiddie porn.



I'm not in support of it, but it's better than the real thing.

I'd rather a pedophile have a child sex doll, than a child. The aim should be to minimize harm, and I don't give a damned how that's accomplished.
41   RC2006   2023 Feb 3, 11:17am  

I think this AI stuff is going to be an out for people that get exposed, they will say it was AI and fake. Eventually we wont be able to tell the difference
43   Eric Holder   2023 Mar 27, 11:47am  

ChatGPT Is a Blurry JPEG of the Web

https://www.newyorker.com/tech/annals-of-technology/chatgpt-is-a-blurry-jpeg-of-the-web

"Imagine that you’re about to lose your access to the Internet forever.
In preparation, you plan to create a compressed copy of all the text on
the Web, so that you can store it on a private server. Unfortunately,
your private server has only one per cent of the space needed; you can’t
use a lossless compression algorithm if you want everything to fit.

"If a compression algorithm is designed to reconstruct text after
ninety-nine per cent of the original has been discarded, we should
expect that significant portions of what it generates will be entirely
fabricated.

This analogy makes even more sense when we remember that a common
technique used by lossy compression algorithms is interpolation—that is,
estimating what’s missing by looking at what’s on either side of the
gap. When an image program is displaying a photo and has to reconstruct
a pixel that was lost during the compression process, it looks at the
nearby pixels and calculates the average. This is what ChatGPT does when
it’s prompted to describe, say, losing a sock in the dryer using the
style of the Declaration of Independence: it is taking two points in
“lexical space” and generating the text that would occupy the location
between them. (“When in the Course of human events, it becomes necessary
for one to separate his garments from their mates, in order to maintain
the cleanliness and order thereof. . . .”) ChatGPT is so good at this
form of interpolation that people find it entertaining: they’ve
discovered a “blur” tool for paragraphs instead of photos, and are
having a blast playing with it."
44   cisTits   2023 Mar 27, 1:56pm  

They also plagiarize a lot, too.
45   Tenpoundbass   2023 Mar 27, 4:28pm  

RC2006 says

I think this AI stuff is going to be an out for people that get exposed, they will say it was AI and fake.


That was my first thought when the CIA was making it, and the MSM was reporting that we should all be careful of videos in the not so distant future, it would be riddled with A.I.
This was before most anyone had even seen their first A.I. generated video. They just seemed so set on setting it up to be the scapegoat when the fake news moniker fails.
46   HeadSet   2023 Mar 27, 5:37pm  

A.I. generated video is just an advanced form of the "trick photography" that has been around for over a century. Like the flying monkeys in "Wizard of Oz" and Audrey Hepburn singing with someone else's voice.
47   Tenpoundbass   2023 Mar 27, 5:51pm  

HeadSet says

just an advanced form of the "trick photography" that has been around for over a century.


I remember when commercials first started using special effects, only they didn't call it special effects. Until that movie FX came out.
My father used to marvel at the Gleem commercial where the White, Green and Red tooth paste on three separate toothbrushes would merge into one. He said "that's one hell of a trick photography!"
48   HeadSet   2023 Mar 27, 6:09pm  

Tenpoundbass says

My father used to marvel at the Gleem commercial where the White, Green and Red tooth paste on three separate toothbrushes would merge into one.

He must have really liked the giant in the washer and the tidy bowl man. How about that camera movement where they made Reagan and Dukakis look the same height during the debate?

« First        Comments 22 - 48 of 48   

Please register to comment:

api   best comments   contact   latest images   memes   one year ago   random   suggestions