WAN Topic? - Principal alleges viral audio recording is AI generated

[so_Kind_]

Original article from Baltimore Banner:

 

(Alleged remarks may be triggering)!

 

https://www.thebaltimorebanner.com/education/k-12-schools/pikesville-high-principal-eric-eiswert-NT7K7N4K6RDEJNL5Z7ULTEG7VY/

 

How the heck do we tell what's real and what isn't when some public figures at smaller scales don't have larger platforms to compare to?

[AbydosOne]

Quote

“You could probably do this type of stuff on a computer,” he said.

I get what he means, but literally read, that is a really stupid quote.

[so_Kind_]

2 hours ago, AbydosOne said:

I get what he means, but literally read, that is a really stupid quote.

Even without the quote, the dialogue highlights that people in important positions are not knowledgeable enough about technology and AI in general - at least to me

[Alex Atkin UK]

12 minutes ago, Mgrooms88 said:

Even without the quote, the dialogue highlights that people in important positions are not knowledgeable enough about technology and AI in general - at least to me

I mean its not like you can magic up audio cloning, its actually quite hard and AFAIK requires very clean source material to train on.  Even then odds are its not going to sound natural unless you pay someone a ton of money to tweak it.

 

There's a big difference between AI audio that is "good enough" for marketing etc, and good enough to actually defame someone.

 

When it comes to normal sentence structure of a person, that's a heck of a lot harder to reproduce than reading a script.

[Flavius Heraclius]

Regardless of which way this one actually went, I'm am not looking forward to the coming future when anybody can plausibly say that recordings/footage of them saying/doing something questionable was faked.

 

I honestly don't know which will be worse: using AI to make people say things that they didn't, or people trying to use the fact that something they said "could have been faked" to convince people that they didn't say it.

EDIT: Particularly because we already know that somebody will usually be much more likely to believe whichever narrative they encounter first on any given issue, regardless of what the future evidence or discussion turns out to be.

[so_Kind_]

12 minutes ago, Flavius Heraclius said:

Regardless of which way this one actually went, I'm am not looking forward to the coming future when anybody can plausibly say that recordings/footage of them saying/doing something questionable was faked.

 

I honestly don't know which will be worse: using AI to make people say things that they didn't, or people trying to use the fact that something they said "could have been faked" to convince people that they didn't say it.

EDIT: Particularly because we already know that somebody will usually be much more likely to believe whichever narrative they encounter first on any given issue, regardless of what the future evidence or discussion turns out to be.

^

This. I'm really not looking forward to the PUBLIC having to come to a consensus on what happened either

[Needfuldoer]

11 hours ago, Flavius Heraclius said:

Regardless of which way this one actually went, I'm am not looking forward to the coming future when anybody can plausibly say that recordings/footage of them saying/doing something questionable was faked.

 

I honestly don't know which will be worse: using AI to make people say things that they didn't, or people trying to use the fact that something they said "could have been faked" to convince people that they didn't say it.

EDIT: Particularly because we already know that somebody will usually be much more likely to believe whichever narrative they encounter first on any given issue, regardless of what the future evidence or discussion turns out to be.

As fun and exciting as all this AI stuff is, it's going to dismantle whatever semblance of objective, consensus reality we've got left.

 

"[Politician I like] would never say that, that has to be fake. I don't believe it! I'm not going to listen to that! I'm going to believe [politician I like]!"

[Holmes108]

12 hours ago, Alex Atkin UK said:

There's a big difference between AI audio that is "good enough" for marketing etc, and good enough to actually defame someone.

 

 

If you're talking about something holding up in court, I'd be there with you. If we're talking about the court "of public opinion", I'm not so sure. Seems pretty easy to (sucessfully) defame someone these days even without a recording.

[Alex Atkin UK]

37 minutes ago, Holmes108 said:

If you're talking about something holding up in court, I'd be there with you. If we're talking about the court "of public opinion", I'm not so sure. Seems pretty easy to (sucessfully) defame someone these days even without a recording.

All bets are off when it comes to public opinion though, you don't even need proof these days.  Just look at how people still believe every word Trump says despite the vast amount of evidence against it.

[Holmes108]

17 minutes ago, Alex Atkin UK said:

All bets are off when it comes to public opinion though, you don't even need proof these days.  Just look at how people still believe every word Trump says despite the vast amount of evidence against it.

 

That's what I mean. It's a scary future either way, to be sure. I don't even know what this means for the justice system as the technology gets better and better going forward.