Live data from Hacker News

They stole my voice with AI

jeffgeerling.com

121–130 of 461 posts

Re: They stole my voice with AI

#121
post #23

Earlier quoted context omitted.

It's going to be an interesting First Amendment question.

Might as well make photoshopping and manual audio tweaking / impersonation illegal, since its the same ballpark, just less effort.

So currently, if I Photoshop a picture of Scarlett Johansen into in an ad for my hot dog stand, that would be unambiguously no problem? Nobody's talking about making AI illegal, but some people seem to think it's a get-out-of-jail-free card for copyright violations, and it's just not.

Re: They stole my voice with AI

#122

Earlier quoted context omitted.

> Theft is undefined for objects with .copy() interface. > Computers replaced computers, now voice acting replaces voice actors. It's incredible what web development does to someone's ability to communicate ideas.

The original meaning of the word computer was a person who calculates. Looks perfectly clear to me.

I have a very funny example of how using outdated words might lead to miscommunication, but unfortunately that particular one could get me banned, so let's just say that I'm a huge fan of picking a version of langauge from one specific time period and sticking to it, instead of having the reader do all the guesswork.

Re: They stole my voice with AI

#123

If LLMs are the ultimate remix machine, then is anyone with a RAG a digital DJ? One can't help but wonder what theft even means any more, when it comes to digital information. With the (lack of) legal precedent, it feels like the wild wild west of intellectual property and copyright law. Like, if even a superstar like Scarlett Johansson can only write a pained letter about OpenAI's hustle to mimic her "Her" persona,…

I assume Jeff wants cease and desist too as this seems more blatant on the surface. Starts a cat and mouse game until they find a variation they feel is different enough to ignore his pleads. Some will use this new clone tech for free publicity hack and others will claim it's still their voice and try to censor it as punishment for targeting them or not doing a bigger deal for the real voice and finding a better one.

Re: They stole my voice with AI

#125

This whole argument rests on the absurd assumption that you can "own" a voice as if it's property. Does this mean people can own the patterns of vibration in air? It's completely nonsensical.

>Does this mean people can own the patterns of vibration in air? It's completely nonsensical.

If you argue similarly, then the whole juridical system is nonsensical, because everything is just particles and waves, and different configurations thereof - not to mention the many protected things that are acts, which are neither particles nor waves, and are completely made up.

I'd say it's desirable to regulate something like this, however nonsensical-seeming, so that we can at least somewhat protect the individuals, and the general well-being of society.

Re: They stole my voice with AI

#126

Earlier quoted context omitted.

> Theft is undefined for objects with .copy() interface. > Computers replaced computers, now voice acting replaces voice actors. It's incredible what web development does to someone's ability to communicate ideas.

The original meaning of the word computer was a person who calculates. Looks perfectly clear to me.

> The original meaning of the word computer was a person who calculates.

Was that ever so?

According to Google Ngrams, "computer" was not really a term until, predictably, around 1950, even though the verb "compute" has been in use throughout history (it comes from the Latin "computare", apparently).

That tells me "computer" was never a person, always a machine.

https://books.google.com/ngrams/graph?content=compute%2Ccomp...

Re: They stole my voice with AI

#127

If LLMs are the ultimate remix machine, then is anyone with a RAG a digital DJ? One can't help but wonder what theft even means any more, when it comes to digital information. With the (lack of) legal precedent, it feels like the wild wild west of intellectual property and copyright law. Like, if even a superstar like Scarlett Johansson can only write a pained letter about OpenAI's hustle to mimic her "Her" persona,…

> feels like the wild wild west of intellectual property and copyright law Copyright seems to always have one or another wild wild west going on. Maybe you are in the wrong place if the world constantly jumps and kicks from under you trying to throw you off?

Anyone that thinks this is completely untrodden ground for copyright should ask an expert to definitely determine if someone's use of something is covered under fair use if it doesn't exactly and clearly satisfy all of the test prongs.

Re: They stole my voice with AI

#128

It's all fun and games until someone produces a recording of somebody else saying something incriminating and it will be used in court. This is the part of AI I hate.

Unless you also hate Image Editors, I don’t really get this point. Preserving forms of evidence isn’t really a primary concern when evaluating new useful technology.

Re: They stole my voice with AI

#129
(Obviously not a lawyer) Overlooking the AI part, isn't it a gross misrepresentation of Jeff's opinion or an unauthorized use of his image? By using his voice, it creates an implicit (fabricated) endorsement for their product and that feels very wrong. I'm sure laws exists to deal with these cases, since way before AI existed.

Re: They stole my voice with AI

#130
post #32

The idea that stolen voice tones are going to matter at all is one of the shortest sighted bits of AI investment - powered by Hollywood "never make anything new" thinking. In about 5 years AI voices will be bespoke and more pleasant to listen to then any real human: they're not limited by vocal cord stress, can be altered at will, and can easily be calibrated by surveying user engagement. Subtly tweaking voice output…

I am not convinced that it will be even 5 years. Have you tested elevenlabs[1]? They offer different voice cloning techniques today, starting from 30 seconds of audio input (sounds somewhat like the cloned voice but definitely not exactly the same) to multiple hours of voice input (sounds like the actual person). In addition, you can adjust the voices with a few parameters or simply create one by defining parameters.…

It's not the instant cloning that's the issue, it's cloning and tweaking - I don't think we quite have the methodologies built yet to optimize it.

But we know it does matter - i.e. there's research which shows a good sound quality on a voice call improves whether people believe what you say[1].

Now in any individual session, you probably can't make particularly big alterations, but imagine say, Google or Amazon shipping a modified voice assistant voice as "the default" with every new speaker box? Whether people ask for the default voice, or change it, would all become data which tells you what people are responding to. And so right there, your new "voice of Google" or "voice of Amazon" you use in other places now becomes informed by wide-scale testing of whether people listen to it.

And that's presuming no one simply runs studies where they stick people in fMRI machines and play them an AI voice recording which they module according to neural feedback till it's "optimal".

[1] https://today.usc.edu/why-we-believe-something-audio-sound-q...

Post reply on HN