“Generative AI is a sack of wet garbage”

[bilby]

In a few years we may be unable to distinguish reality from fiction.

For many centuries, many people have been unable to distinguish reality from fiction. This is just more of the same.

Most of the fakes out there aren't high quality--they don't need to be to fool the average person inclined to support the position.

It's always the enemy you can't see that's the biggest threat.

 

[unapologetic]

Now Kamala Harris is AI generated. The giveaway is the car passing beneath the window. Same car over and over.
Congress UNANIMOUSLY Votes to Strip Trump of All Presidential Authority! (misleading title)

I've previously seen another vid of her from this room, with random cars, and I was fooled.

 

[Swammerdami]

George Will gives a long presentation -- it seems that Melania is preparing to testify against her husband about his felony tax frauds. Since she signed the joint tax returns, her skin is on the line too. She's making an immunity deal which the Supreme Court just OK'ed. The face and voice seem to be those of the real George Will.

TOTAL FICTION. One give-away is that at one point the AI(?) blunders and refers to her "former husband." Sometimes a hard-to-see "altered or synthetic content" label appears briefly, but I don't even know how to get it to reappear.

What's the purpose of creating such a thing? just clicks at a fraction of a penny each? Some teenager honing his skills? How can we trust anything now?

 

[bigfield]

How can we trust anything now?

THe same way we always have. We reserve our trust for people and institutions that have earned it, not some random shitstirrers on social media.

 

[Swammerdami]

How can we trust anything now?

THe same way we always have. We reserve our trust for people and institutions that have earned it, not some random shitstirrers on social media.

I've watched George Will and listened to him. I thought this was him.

Looking at poster names mayn't help you. Do the character in the three strings following look the same to you?

oooooooooo == οοοοοοοοοοο == oοoοoοoοoοoο

 

[bigfield]

How can we trust anything now?

THe same way we always have. We reserve our trust for people and institutions that have earned it, not some random shitstirrers on social media.

I've watched George Will and listened to him. I thought this was him.

But you didn't get the video from a trustworthy source like a reputable news outlet.

 

[bigfield]

Looking at poster names mayn't help you. Do the character in the three strings following look the same to you?

oooooooooo == οοοοοοοοοοο == oοoοoοoοoοoο

Sorry not sure what you're getting at?

 

[Swammerdami]

Looking at poster names mayn't help you. Do the character in the three strings following look the same to you?

oooooooooo == οοοοοοοοοοο == oοoοoοoοoοoο

Sorry not sure what you're getting at?

Suppose somebody thinks that The Boston Globe is a trustworthy source, and they see something that appears to be from "The Boston Globe." How can they be sure it's from The Boston Globe? Even if there's some authority that prevents an impostor from using that name, maybe they use the name The Bοstοn Glοbe instead. Fooled you? Use your browser's Ctrl-F facility to search for "Boston" or "Globe."

This The Bοstοn Glοbe will NOT show up in your search. It is pretending to be a name that it's not. I am NOT suggesting that the little trick I used here to produce a fake The Bοstοn Glοbe is an important way to deceive. There are many other ways.

Note that I wasn't deceived by someone usurping just the name "George Will." The face and voice REALLY appeared to be him. I caught on because the allegations were too startling, too unbelievable. Had the content creator set his target lower it could be more convincing.

And even if we stipulate that bigfield is 100 times as smart as gullible Swammi, and that bigfield could never fall for such fakery, there are millions of Americans at least as stupid as Swammi. If you think that fake propaganda does not harm YOU as long as YOU personally are too smart to fall for it, I believe you are mistaken. YOU didn't vote for Trump, but he is still your President.

 

[bilby]

And even if we stipulate that bigfield is 100 times as smart as gullible Swammi, and that bigfield could never fall for such fakery, there are millions of Americans at least as stupid as Swammi. If you think that fake propaganda does not harm YOU as long as YOU personally are too smart to fall for it, I believe you are mistaken. YOU didn't vote for Trump, but he is still your President.

I don't disagree with your broad argument here, but will point out that bigfield is neither a US citizen nor a US resident, and that Trump is not, therefore, his President.

Unless, of course, his posting history here is all a fake generated by propagandists with an extraordinary talent for subtlety and detail.

 

[bigfield]

Suppose somebody thinks that The Boston Globe is a trustworthy source, and they see something that appears to be from "The Boston Globe." How can they be sure it's from The Boston Globe? Even if there's some authority that prevents an impostor from using that name, maybe they use the name The Bοstοn Glοbe instead. Fooled you? Use your browser's Ctrl-F facility to search for "Boston" or "Globe."

The means vary by medium (Youtube, TV, print media etc.) but I'm pretty sure you and I can confirm the authenticity of our sources.

I've been fooled by plain old satire, let alone deepfakes. Anyone can get fooled when we aren't being critical.

Note that I wasn't deceived by someone usurping just the name "George Will." The face and voice REALLY appeared to be him. I caught on because the allegations were too startling, too unbelievable. Had the content creator set his target lower it could be more convincing.

Maybe this is a sign that you need to calibrate the way you critically consume online media?

 

[ideologyhunter]

This week's New Yorker has a piece about some film makers who are at present working on the Orson Welles picture The Magnificent Ambersons (1942), to create a restored version. RKO butchered the film and cut something like 40 minutes out of it, then junked the cut footage. The AI team is working with the shooting script and continuity (description of camera work) to recreate lost scenes. For Joseph Cotten's scenes, they will have a live actor go through the physical motion and dialogue, then create a likeness of Cotten and his voice and superimpose it over the 2026 actor. And so on, with the other original cast members. The Welles estate, which is administered by his daughter, is conflicted about the whole thing. (At first the daughter reacted with indignation, but she has come to consider that the AI team is made up of Welles devotees, who think he's the top film maker of all time.)
I'm a huge buff of classic film & I'm not sure how I feel about this. Definitely not sure I'd want to watch it; my attention would be taken up with trying to spot the monkey-gland portions.
When I watch The Irishman, I find it difficult to watch DeNiro especially in that cast, with the 'rejuvenation magic' they worked on him. It makes him look like an Asian man of indeterminate age with limited facial mobility. Pacino looks natural, Pesci a little on the squinky side. And that film was released seven years ago, I believe. The tech is developing at warp speed.
The mischief with politics is already here, and disruptive. How long will it be before candidates will have to end each ad with, "I'm <your name here> and I approved of this message, plus, I am not a bot, honest injun"?

 

[Swammerdami]

Note that I wasn't deceived by someone usurping just the name "George Will." The face and voice REALLY appeared to be him. I caught on because the allegations were too startling, too unbelievable. Had the content creator set his target lower it could be more convincing.

Maybe this is a sign that you need to calibrate the way you critically consume online media?

You write as though Swammi were the problem. I don't even vote^*, and anyway have already outlived the "three-score and ten" default. I'm likely to miss out on future excitement.

But millions -- make that billions -- will be victimized by the increasing amounts of "wet garbage" arising from generative AI. They will be victimized whether more or less critical than I. In many cases they WILL get the clear message that media cannot be trusted; they will then distrust even the GOOD media, if there still is such a thing.

Maybe your genius-level critical ability will save the world! I sure hope so.

* -

Spoiler: Why I don't vote

I can vote legally only in California where elections are typically Democrat -vs- Democrat. If I tried to vote there's a fair chance that California -- a place I've not visited for decades -- would assess personal income taxes. They've done that before.

 

[bigfield]

You write as though Swammi were the problem.

I thought you had a problem? Am I wrong?

 

[unapologetic]

In the past month I've seeen YT vids with AI versions of, Kamala Harris, Rachel Maddow, Jimmy Kimmel, and John Oliver. (misspellings would be mine)
The AI John Oliver was less funny and less animated than the real John.
One tell: he said a lady was "a four, one year old." instead of 41. A mistake no real human would ever make, even reading a typo off a teleprompter.
If you make your living on TV, can't you sue for this. It's not satire.

 

[ZiprHead]

In the past month I've seeen YT vids with AI versions of, Kamala Harris, Rachel Maddow, Jimmy Kimmel, and John Oliver. (misspellings would be mine)
The AI John Oliver was less funny and less animated than the real John.
One tell: he said a lady was "a four, one year old." instead of 41. A mistake no real human would ever make, even reading a typo off a teleprompter.
If you make your living on TV, can't you sue for this. It's not satire.

That's a good question.

I think a cease and desist letter to both the poster and Google/Meta/whatever they're calling it threatening lawsuits would help. Whether those lawsuits would be successful is a question I cannot answer but I think just the threat would help alleviate the problem. Would it be worth the time and money for the platform where the vids are being posted to defend? I don't think so.

 

[thebeave]

This headline and article from The Guardian (aka Teh Gruniand) sounds like AI slop to me.

It’s not easy being green: Trump’s botched reflecting pool becomes 2,028ft metaphor

Presumably, the 2,028 is referencing the next presidential election year. I have seen AI content frequently use commas when it comes to writing and pronouncing years, which makes me think this is all raw AI slop. Plus, what's with "ft"? There is no mention of length in the article. You'd think they could at least have a human spend a few minutes proofreading?

 

[Don2 (Don1 Revised)]

This headline and article from The Guardian (aka Teh Gruniand) sounds like AI slop to me.

It’s not easy being green: Trump’s botched reflecting pool becomes 2,028ft metaphor

Presumably, the 2,028 is referencing the next presidential election year.

Perhaps. Also, a metaphor has two things: one being (compared directly) to another, as if it were the thing. So if it means the presidential year, it may also mean something else.

I have seen AI content frequently use commas when it comes to writing and pronouncing years, which makes me think this is all raw AI slop. Plus, what's with "ft"?

You perhaps answered your own question. The length of the reflecting pool is 2028 ft, comma is optional.

There is no mention of length in the article.

There is in the title itself. Since the length of the pool is 2028ft, the pool can be called a 2028ft metaphor without mentioning the 2028ft in the article. Does the article mention things like grifting, corruption, lies, exaggerations, and blame?

You'd think they could at least have a human spend a few minutes proofreading?

I think you might need a stronger proof of AI involvement in the title.

That said, so much AI use _does_ begin to make non-AI use suspect. It's harder to trust anything these days.

 

[thebeave]

This headline and article from The Guardian (aka Teh Gruniand) sounds like AI slop to me.

It’s not easy being green: Trump’s botched reflecting pool becomes 2,028ft metaphor

Presumably, the 2,028 is referencing the next presidential election year.

Perhaps. Also, a metaphor has two things: one being (compared directly) to another, as if it were the thing. So if it means the presidential year, it may also mean something else.

I have seen AI content frequently use commas when it comes to writing and pronouncing years, which makes me think this is all raw AI slop. Plus, what's with "ft"?

You perhaps answered your own question. The length of the reflecting pool is 2028 ft, comma is optional.

There is no mention of length in the article.

There is in the title itself. Since the length of the pool is 2028ft, the pool can be called a 2028ft metaphor without mentioning the 2028ft in the article. Does the article mention things like grifting, corruption, lies, exaggerations, and blame?

You'd think they could at least have a human spend a few minutes proofreading?

I think you might need a stronger proof of AI involvement in the title.

That said, so much AI use _does_ begin to make non-AI use suspect. It's harder to trust anything these days.

Is it supposed to be common knowledge that the pool is 2028 ft long? If so, then the title could somewhat make sense. But this is first time I've heard it's length mentioned.

 

[Don2 (Don1 Revised)]

This headline and article from The Guardian (aka Teh Gruniand) sounds like AI slop to me.

It’s not easy being green: Trump’s botched reflecting pool becomes 2,028ft metaphor

Presumably, the 2,028 is referencing the next presidential election year.

Perhaps. Also, a metaphor has two things: one being (compared directly) to another, as if it were the thing. So if it means the presidential year, it may also mean something else.

I have seen AI content frequently use commas when it comes to writing and pronouncing years, which makes me think this is all raw AI slop. Plus, what's with "ft"?

You perhaps answered your own question. The length of the reflecting pool is 2028 ft, comma is optional.

There is no mention of length in the article.

There is in the title itself. Since the length of the pool is 2028ft, the pool can be called a 2028ft metaphor without mentioning the 2028ft in the article. Does the article mention things like grifting, corruption, lies, exaggerations, and blame?

You'd think they could at least have a human spend a few minutes proofreading?

I think you might need a stronger proof of AI involvement in the title.

That said, so much AI use _does_ begin to make non-AI use suspect. It's harder to trust anything these days.

Is it supposed to be common knowledge that the pool is 2028 ft long? If so, then the title could somewhat make sense. But this is first time I've heard it's length mentioned.

How do you deduce it ought to be common knowledge? In my reading of the title, it comes across as an implication. I figured it out from reading the title, then confirming. I confirmed because you offered an alternate explanation.