Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: moviePig <nobody@nowhere.com>
Newsgroups: rec.arts.tv
Subject: Re: OT: The AIs have it...
Date: Sun, 2 Mar 2025 17:53:29 -0500
Organization: A noiseless patient Spider
Lines: 91
Message-ID: <vq2nh9$vm7q$1@dont-email.me>
References: <vpnugk$2nkhb$1@dont-email.me>
 <0001HW.2D7416B003C8F76A30A6CD38F@news.giganews.com>
 <vq0n8g$ksbf$2@dont-email.me> <vq23q4$rvpu$1@dont-email.me>
 <vq2it3$umv9$4@dont-email.me>
Reply-To: nobody@nowhere.com
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Sun, 02 Mar 2025 23:53:30 +0100 (CET)
Injection-Info: dont-email.me; posting-host="2cd3803a06b4684e4a2ba8a1249dae23";
	logging-data="1038586"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX18OXMZrxwyplE14cc01kzCSfLX7h8+4Zdk="
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:miXFSRaBG8zi0Xt50r/nHAsu1cQ=
Content-Language: en-US
In-Reply-To: <vq2it3$umv9$4@dont-email.me>
Bytes: 4999

On 3/2/2025 4:34 PM, BTR1701 wrote:
> On Mar 2, 2025 at 9:16:51 AM PST, "moviePig" <nobody@nowhere.com> wrote:
> 
>> On 3/1/2025 11:36 PM, BTR1701 wrote:
>>>   On Mar 1, 2025 at 8:31:44 PM PST, "Pluted Pup" <plutedpup@outlook.com>
>>> wrote:
>>>   
>>>>   On Sat, 01 Mar 2025 19:53:55 -0800, BTR1701 wrote:
>>>>
>>>>>     On Mar 1, 2025 at 4:51:12 PM PST, "Pluted Pup"<plutedpup@outlook.com>
>>>>>   wrote:
>>>>>
>>>>>     > On Wed, 26 Feb 2025 16:34:00 -0800, BTR1701 wrote:
>>>>>     >
>>>>>     > > On Feb 26, 2025 at 3:06:45 PM PST, "Alan Smithee"<alms@last.inc>  wrote:
>>>>>     > >
>>>>>     > > > 1,000 artists release a silent album to protest AI taking their
>>>>>>>> works...
>>>>>     > > >
>>>>>     > > >
>>>>>     > > >
>>>>>>>>
>>>>>>>> https://www.techspot.com/news/106909-over-1000-musicians-release-silent-album-protest-ai.html
>>>>>     > >
>>>>>     > > I've never understood the claim that training AI systems on books,
>>>>>>> music,
>>>>>     > > etc.
>>>>>     > > is a copyright violation in the first place.
>>>>>     > >
>>>>>     > > The AI isn't making an unauthorized copy of the work. It's reading (or
>>>>>     > > listening to ) the work and learning from it. This isn't any different
>>>>>>>   than
>>>>>     > > a
>>>>>     > > human being reading a book and learning from it.
>>>>>     > >
>>>>>     > > Some have said, well, the AI makes a copy of the work in its brain while
>>>>>     > > it's
>>>>>     > > learning but the same can be said of a human. Why is one a (supposed)
>>>>>     > > copyright violation but the other is not?
>>>>>     >
>>>>>     > You use your brain to violate copyright law or tell a computer
>>>>>     > to violate copyright law and you say the computer user should get a free
>>>>>     > pass?
>>>>>
>>>>>     No, I'm saying that a human reading a book with her brain DOESN'T violate
>>>>>     copyright law, so why should a computer reading a book with its brain
>>>>>   become a
>>>>>     violation?
>>>>
>>>>   No, I am saying that someone committing copyright fraud with computers
>>>>   shouldn't be exonerated while only those using their own brain to
>>>>   commit copyright fraud should be prosecuted.
>>>   
>>>   I have no idea what you're talking about. No one's being prosecuted for
>>>   committing "copyright fraud' (whatever that is) with their brains.
>>>   
>>>>   That's mindless.
>>>   
>>>   Indeed.
>>
>> While you can't copyright ideas, you can copyright *expressions* of
>> ideas.  When you read a book and understand its ideas, you can then
>> freely voice them from your understanding.  If, however, you *don't*
>> understand the ideas, and instead merely parrot their expression from
>> the book, you violate copyright.
> 
> No, you violate copyright when you copy a work without permission. Based on
> the language in the statute and the court cases interpreting it over the last
> 40 years or so, the AI models in no way are creating legal copies of the works
> in question when they amalgamate the knowledge and facts found in millions of
> books to answer people's questions on the internet.
> 
>> AIs, as yet, have no claim of such
>> understanding, and instead rely on (sophisticated) parroting.

Afaics, 'permission' isn't at issue here, and thus is irrelevant.  But 
what is "the amalgamation of knowledge"?  If it's a level of 
understanding that effectively discards the original rendering, then 
there's no violation.  If, otoh, it's the mere memorization of phrases, 
*without* any real understanding, then that violates copyright 
(regardless of judicial miscarriage to the contrary).  For example:

    "Booze works faster than chocolate"

wouldn't violate copyright, whereas:

    "Candy is dandy, but liquor is quicker"

would.